Multi-start downhill simplex method for spatio-temporal source localization in magnetoencephalography

Huang, Mingxiong; Aine, Cheryl J.; Supek, Selma; Best, Elaine; Ranken, Doug; Flynn, E.R.

doi:10.1016/s0168-5597(97)00091-9

Cited by 124 publications

(102 citation statements)

References 21 publications

Supporting

Mentioning

102

Contrasting

Order By: Relevance

“…The time interval of 20-870 ms poststimulus was analyzed for all task conditions. We developed and used the automated Multistart Downhill Simplex Method for determining source locations, strengths, and magnitudes (Huang et al, 1998;Aine et al, 2000). Nelder and Mead's (1965) downhill simplex search was utilized along with the Sarvas (1987) formula for forward calculations.…”

Section: Source Localization and Statistical Methodsmentioning

confidence: 99%

Aging: Compensation or maturation?

Aine

Woodruff²,

Knoefel

et al. 2006

NeuroImage

View full text Add to dashboard Cite

Neuroimaging studies of healthy aging often reveal differences in neural activation patterns between young and elderly groups for episodic memory tasks, even though there are no differences in behavioral performance. One explanation typically offered is that the elderly compensate for their memory deficiencies through the recruitment of additional prefrontal regions. The present study of healthy aging compared magnetoencephalographic (MEG) time-courses localized to specific cortical regions in two groups of subjects (20-29 years and ≥65 years) during a visual delayed-match-to-sample (DMS) task. MR morphometrics and neuropsychological test results were also examined with the hope of providing insight into the nature of the age-related differences. The behavioral results indicated no differences in performance between young and elderly groups. Although there was a main effect of age on the latency of the initial peak in primary/secondary visual cortex, these longer latencies were not correlated with the performance of elderly on the DMS task. The lateral occipital gyrus (LOG) revealed qualitatively different patterns of activity for the two age groups corroborated by neuropsychological test results. Morphometric results for the young versus elderly groups revealed less white (WM) and gray matter (GM) volumes in the frontal lobes of the elderly. When a group of middle-aged subjects (33-43 years) was included in the morphometric analyses, the middle-aged subjects revealed statistically greater WM volumes in frontal and parietal cortex suggesting immature WM tracts in the young. Perhaps our elderly utilized a different strategy compared to the young due to the different brain maturation levels of these groups.

show abstract

Section: Source Localization and Statistical Methodsmentioning

confidence: 99%

Aging: Compensation or maturation?

Aine

Woodruff²,

Knoefel

et al. 2006

NeuroImage

View full text Add to dashboard Cite

show abstract

“…Instead of using an informed initialization that anticipates some best solution, we run the algorithm several times (using different random initializations) and select the best solution: a similar multi-start procedure has been described by Huang et al (1998). In our experience, to find a good solution, one should at least run four iterations for single dipole models.…”

Section: Initialization and Convergencementioning

confidence: 99%

Variational Bayesian inversion of the equivalent current dipole model in EEG/MEG

et al. 2008

View full text Add to dashboard Cite

In magneto-and electroencephalography (M/EEG), spatial modelling of sensor data is necessary to make inferences about underlying brain activity. Most source reconstruction techniques belong to one of two approaches: point source models, which explain the data with a small number of equivalent current dipoles and distributed source or imaging models, which use thousands of dipoles. Much methodological research has been devoted to developing sophisticated Bayesian source imaging inversion schemes, while dipoles have received less such attention. Dipole models have their advantages; they are often appropriate summaries of evoked responses or helpful first approximations. Here, we propose a variational Bayesian algorithm that enables the fast Bayesian inversion of dipole models. The approach allows for specification of priors on all the model parameters. The posterior distributions can be used to form Bayesian confidence intervals for interesting parameters, like dipole locations. Furthermore, competing models (e.g., models with different numbers of dipoles) can be compared using their evidence or marginal likelihood. Using synthetic data, we found the scheme provides accurate dipole localizations. We illustrate the advantage of our Bayesian scheme, using a multi-subject EEG auditory study, where we compare competing models for the generation of the N100 component. © 2007 Elsevier Inc. All rights reserved. Keywords: EEG; MEG; Equivalent current dipole; Variational Bayes IntroductionThe analysis of evoked responses using magneto-and electroencephalography (M/EEG) can proceed in several ways. If one is interested in inferring the locations of M/EEG generators within brain space, one has to solve the inverse spatial problem (Baillet et al., 2001). There are two main approaches to estimating sources from observed sensor data. The first assumes that sensor data can be explained by a small set of equivalent current dipoles. The inversion of this model amounts to a nonlinear optimization problem, because the forward model is nonlinear in dipole location (Mosher et al., 1992). Recently, the source reconstruction problem has been addressed by placing many dipoles in brain space, and using constraints on the solution to make it unique; for example (Baillet and Garnero, 1997;Mattout et al., 2006;Phillips et al., 2005). This approach is attractive, because it produces images of brain activity comparable to other imaging modalities and it eschews subjective constraints on the inversion. For imaging solutions, most constraints can be motivated by anatomical and physiological arguments, e.g., smoothness constraints and approximate location priors, based on regional activity in functional magnetic resonance imaging (fMRI). Traditional few-dipole solutions, however, are usually regarded as depending too much on user-specified modelling decisions; like the number of dipoles and their initial locations. Mathematically, it can be argued that the inversion of dipole models is a harder problem than inversion of distributed models, becaus...

show abstract

“…For the large dipole grid used in our median-nerve MEG data analysis with 16,137 dipole sites, the computational time for any of the beamformers was less than 50 sec using MATLAB (Math Works Inc., Massachusetts, USA) on a 600 MHz Pentium III PC. By comparison, a multiple dipole solution using nonlinear global optimization algorithms usually takes a couple of hours for this type of data (Huang et al 1998;Uutela et al 1998). …”

Section: Source Orientations and The Order Of Covariance Matrixmentioning

confidence: 99%

“…Unlike multiple dipole fitting where the number of dipoles has to be determined in advance (Huang et al 1998), the beamformer requires no a priori assumptions about the number of sources to model (Van Veen et al 1997). In addition, the dipole approach which depends on assumption of a point source may not be able to accurately describe the underlying neuronal sources with large extents, whereas the beamformer has the capability to handle such sources (Van Veen et al 1997).…”

Section: Introductionmentioning

confidence: 99%

Commonalities and Differences Among Vectorized Beamformers in Electromagnetic Source Imaging

Huang¹,

Shih²,

Lee³

et al. 2003

Brain Topogr

128

117

View full text Add to dashboard Cite

Summary: A number of beamformers have been introduced to localize neuronal activity using magnetoencephalography (MEG) and electroencephalography (EEG). However, currently available information about the major aspects of existing beamformers is incomplete. In the present study, detailed analyses are performed to study the commonalities and differences among vectorized versions of existing beamformers in both theory and practice. In addition, a novel beamformer based on higher-order covariance analysis is introduced. Theoretical formulas are provided on all major aspects of each beamformer; to examine their performance, computer simulations with different levels of correlation and signal-to-noise ratio are studied. Then, an empirical data set of human MEG median-nerve responses with a large number of neuronal generators is analyzed using the different beamformers. The results show substantial differences among existing MEG/EEG beamformers in their ways of describing the spatial map of neuronal activity. Differences in performance are observed among existing beamformers in terms of their spatial resolution, false-positive background activity, and robustness to highly correlated signals. Superior performance is obtained using our novel beamformer with higher-order covariance analysis in simulated data. Excellent agreement is also found between the results of our beamformer and the known neurophysiology of the median-nerve MEG response.

show abstract

Multi-start downhill simplex method for spatio-temporal source localization in magnetoencephalography

Cited by 124 publications

References 21 publications

Aging: Compensation or maturation?

Aging: Compensation or maturation?

Variational Bayesian inversion of the equivalent current dipole model in EEG/MEG

Commonalities and Differences Among Vectorized Beamformers in Electromagnetic Source Imaging

Contact Info

Product

Resources

About