@TECHREPORT\{IMM2006-04144,
    author       = "M. M{\o}rup and L. K. Hansen and J. Parnas and S. M. Arnfred",
    title        = "Decomposing the time-frequency representation of {EEG} using non-negative matrix and multi-way factorization",
    year         = "2006",
    number       = "",
    series       = "",
    institution  = "",
    address      = "",
    type         = "",
    url          = "http://www2.compute.dtu.dk/pubdb/pubs/4144-full.html",
    abstract     = "We demonstrate how non-negative matrix factorization (NMF) can be used to decompose the inter trial phase coherence (ITPC) of multi-channel {EEG} to yield a unique decomposition of time-frequency signatures present in various degrees in the recording channels. The {NMF} optimization is easily generalized to a parallel factor (PARAFAC) model to form a non-negative multi-way factorization (NMWF). While the {NMF} can examine subject specific activities the {NMWF} can effectively extract the most similar activities across subjects and or conditions. The methods are tested on a proprioceptive stimulus consisting of a weight change in a handheld load. While somatosensory gamma oscillations have previously only been evoked by electrical stimuli we hypothesized that a natural proprioceptive stimulus also would be able to evoke gamma oscillations. {ITPC} maxima were determined by visual inspection and these results were compared to the {NMF} and {NMWF} decompositions. Agreement between the results of the visual pattern inspection and the mathematical decompositions was satisfactory showing two significant coherent activities; the predicted 40Hz activity 60 ms after stimulus onset in the frontal-parietal region contralateral to stimulus side and additionally an unexpected 20Hz activity slightly lateralized in the frontal central region. Consequently, also proprioceptive stimuli are able to elicit evoked gamma activity."
}