Structure-seeking multilinear methods for the analysis of fMRI data

Andersen, Anders H.; Rayens, William S.

doi:10.1016/j.neuroimage.2004.02.026

Cited by 93 publications

(133 citation statements)

References 59 publications

Supporting

Mentioning

132

Contrasting

Order By: Relevance

“…Presently, we extended the CP model to include shifts over one mode. The model can be extended to include shifts over additional modes and can also be generalized to data of more modalities than three which naturally arises for instance when including modes such as subjects, conditions or runs (Andersen and Rayens, 2004). Presently, the focus was set on neuroscience data such as EEG and fMRI, however, the model should be readily applicable to other types of neuroimaging data such as magnetoencephalography (MEG) and positron emission tomography (PET).…”

Section: Discussionmentioning

confidence: 99%

“…Notice that the application of CP to EEG was already suggested in the original paper on CP (Harshman, 1970) and was later reinvented in Möcks (1988) under the name topographic component analysis. In Andersen and Rayens (2004) it was further demonstrated how the CP model is useful in the analysis of neuroimaging data such as fMRI (Andersen and Rayens, 2004). Additional applications of multilinear (also called multiway) modeling in EEG and fMRI include (Möcks, 1988;Field and Graupe, 1991;Wang et al, 2000;Beckmann and Smith, 2005;Miwakeichi et al, 2004;Mørup et al, 2006;De Vos et al, 2007;Acar et al, 2007;Dyrholm et al, 2007).…”

Section: Introductionmentioning

confidence: 99%

“…Our modeling of delays is further motivated by a number of papers that explain how degenerate solutions might be caused by component delays (Field and Graupe, 1991;Andersen and Rayens, 2004;Harshman et al, 2003a;Hong and Harshman, 2003). Indeed, if shifts are causing CP-degeneracy, then it is more natural to extend the CP model to accomodate shifts rather than resorting to orthogonality or independence constraints that may not be physiologically justified.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Shift-invariant multilinear decomposition of neuroimaging data

et al. 2008

View full text Add to dashboard Cite

We present an algorithm for multilinear decomposition that allows for arbitrary shifts along one modality. The method is applied to neural activity arranged in the three modalities space, time, and trial. Thus, the algorithm models neural activity as a linear superposition of components with a fixed time course that may vary across either trials or space in its overall intensity or latency. Its utility is demonstrated on simulated data as well as actual EEG, and fMRI data. This work shows how pseudo-multilinear decompositions of multiway data can successfully cope with variable latencies in data derived from neural activitya problem that has caused degenerate solutions especially in modeling neuroimaging data with instantaneous multilinear decompositions. Our algorithm is available for download at www.erpwavelab.org.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Shift-invariant multilinear decomposition of neuroimaging data

et al. 2008

View full text Add to dashboard Cite

show abstract

“…Let A I×D , B J×D and C K×D×T be (hyper-)matrices holding the profiles defined in the convCP model in equation (1). In the following C, B andC,B will denote the same (hyper-)matrix in the time and frequency domain respectively, i.e.…”

Section: The Convcp Modelmentioning

confidence: 99%

“…The application of CP to EEG was first suggested in [11] and was later reinvented in [21] under the name topographic component analysis. In [1] it was further demonstrated how the CP model is useful in the analysis of fMRI.…”

Section: Introductionmentioning

confidence: 99%

Modeling latency and shape changes in trial based neuroimaging data

Mørup

Hansen

Madsen

2011

2011 Conference Record of the Forty Fifth Asilomar Conference on Signals, Systems and Computers (ASILOMAR)

View full text Add to dashboard Cite

Abstract-To overcome poor signal-to-noise ratios in neuroimaging, data sets are often acquired over repeated trials that form a three-way array of space×time×trials. As neuroimaging data contain multiple inter-mixed signal components blind signal separation and decomposition methods are frequently invoked for exploratory analysis and as a preprocessing step for signal detection. Most previous component analyses have avoided working directly with the tri-linear structure, but resorted to bi-linear models such as ICA, PCA, and NMF. Multi-linear decomposition can exploit consistency over trials and contrary to bi-linear decomposition render unique representations without additional constraints. However, they can degenerate if data does not comply with the given multi-linear structure, e.g., due to time-delays. Here we extend multi-linear decomposition to account for general temporal modeling within a convolutional representation. We demonstrate how this alleviates degeneracy and helps to extract physiologically plausible components. The resulting convolutive multi-linear decomposition can model realistic trial variability as demonstrated in EEG and fMRI data.

show abstract