Independent Components of Magnetoencephalography: Single-Trial Response Onset Times

Abstract

We recently demonstrated that second-order blind identification (SOBI), an independent component analysis (ICA) method, can separate the mixture of neuronal and noise signals in magnetoencephalographic (MEG) data into neuroanatomically and neurophysiologically meaningful components. When the neuronal signals had relatively higher trial-to-trial variability, SOBI offered a particular advantage in identifying and localizing neuronal source activations with increased source detectability (A. C. Tang et al., 2002, Neural Comput. 14, 1827-1858). Here, we explore the utility of SOBI in the analysis of temporal aspects of neuromagnetic signals from MEG data. From SOBI components, we were able to measure single-trial response onset times of neuronal populations in visual, auditory, and somatosensory modalities during cognitive and sensory activation tasks, with a detection rate as high as 96% under optimal conditions. Comparing the SOBI-aided detection results with those obtained directly from the sensors, we found that with SOBI preprocessing, we were able to measure, among a greater proportion of trials, single-trial response onset times that are above background neuronal activity. We suggest that SOBI ICA can improve our current capability in measuring single-trial responses from human subjects using the noninvasive brain imaging method MEG.

EEG filtering based on blind source separation (BSS) for early detection of Alzheimer's disease

“…It provides similar decomposition as the well known and popular SOBI algorithms (Belouchrani et al, 1997;Tang et al, 2002). AMUSE algorithm uses simple principles that the estimated components should be spatiotemporally decorrelated and be less complex (i.e.…”

Fast robust subject‐independent magnetoencephalographic source localization using an artificial neural network

“…We see that the primary visual sources are localized more consistently than are the secondary visual sources, across all four tasks. The secondary sources also had more variable stimulus-locked average time courses [Tang and Pearlmutter, 2003]. It is noticeable that somatosensory sources in the right hemisphere are localized poorly by the MLP, but well localized by the hybrid method.…”

“…One such technique that we have used is blind source separation (BSS). BSS has been shown to segregate neuronal from nonneuronal signals, and neuronal signals from each other, in both EEG [Jung et al, 2000a, b;Makeig et al, 1997Makeig et al, , 1999 and MEG [Cao et al, 2000;Tang et al, 2000aTang et al, ,b, 2002Tang and Pearlmutter, 2003;Vigário et al, 1998Vigário et al, , 1999Vigário et al, , 2000Wü bbeler et al, 2000;Ziehe et al, 2000]. For these reasons, despite its limitations, it seems feasible to use the localizer proposed above as a stage in a practical robust real-time MEG processing pipeline.…”

Applying Differentially Variable Component Analysis (dVCA) to Event-related Potentials

“…These algorithms have been useful in identifying sources in EEG and MEG signals using both ensemble-averaged data (Makeig et al, 1997;Sarela et al, 1998; Vighrio et al, 1999) and single trials (Jung et al, 1999;Cao et al, 2000;Makeig et al, 2002;Tang et al, 2002). However, with the exception of SOBI, the general assumption that the ERP sources are independent is physiologically implausible.…”