Weighted Blind Source Separation Can Decompose the Frequency Mismatch Response by Deviant Concatenation: An MEG Study.

The mismatch response (MMR) is thought to be a neurophysiological measure of novel auditory detection that could serve as a translational biomarker of various neurological diseases. When recorded with electroencephalography (EEG) or magnetoencephalography (MEG), the MMR is traditionally extracted by subtracting the event-related potential/field (ERP/ERF) elicited in response to "deviant" sounds that occur randomly within a train of repetitive "standard" sounds. However, there are several problems with such a subtraction, which include increased noise and the neural adaptation problem. On the basis of the original theory underlying MMR (i.e., the memory-comparison process), the MMR should be present only in deviant epochs. Therefore, we proposed a novel method called weighted-BSS T/k, which uses only the deviant response to derive the MMR. Deviant concatenation and weight assignment are the primary procedures of weighted-BSS T/k, which maximize the benefits of time-delayed correlation. We hypothesized that this novel weighted-BSS T/k method highlights responses related to the detection of the deviant stimulus and is more sensitive than independent component analysis (ICA). To test this hypothesis and the validity and efficacy of the weighted-BSS T/k in comparison with ICA (infomax), we evaluated the methods in 12 healthy adults. Auditory stimuli were presented at a constant rate of 2 Hz. Frequency MMRs at a sensor level were obtained from the bilateral temporal lobes with the subtraction approach at 96-276 ms (the MMR time range), defined based on spatio-temporal cluster permutation analysis. In the application of the weighted-BSS T/k, the deviant responses were given a constant weight using a rectangular window on the MMR time range. The ERF elicited by the weighted deviant responses demonstrated one or a few dominant components representing the MMR that fitted well with that of the sensor space analysis using the conventional subtraction approach. In contrast, infomax or weighted-infomax revealed many minor or pseudo components as constituents of the MMR. Our single-trial, contrast-free approach may assist in using the MMR in basic and clinical research, and it opens a new and potentially useful way to analyze event-related MEG/EEG data.

A novel time-delayed correlation method decomposes mismatch response without using subtraction.

The mismatch response (MMR) is thought to be a neurophysiological measure of novel auditory detection that could serve as a translational biomarker of various neurological diseases. When recorded with electroencephalography (EEG) or magnetoencephalography (MEG), the MMR is traditionally extracted by subtracting the event-related potential/field (ERP/ERF) elicited in response to "deviant" sounds that occur randomly within a train of repetitive "standard" sounds. To overcome the limitations of this subtraction procedure, we propose a novel method which we call weighted-BSST/k, which uses only the deviant response to derive the MMR. We hypothesized that this novel weighted-BSST/k method highlights responses related to the detection of the deviant stimulus and is more sensitive than independent component analysis (ICA). To test this hypothesis and the validity and efficacy of the weighted-BSST/k in comparison with ICA (infomax), we evaluated the methods in 12 healthy adults. Auditory stimuli were presented at a constant rate of 2 Hz. Frequency MMRs at a sensor level were obtained from the bilateral temporal lobes with the subtraction approach at 96-276 ms (the MMR time range), defined on the basis of spatio-temporal cluster permutation analysis. In the application of the weighted-BSST/k, the deviant responses were given a constant weight on the MMR time range. The ERF elicited by the weighted deviant responses demonstrated one or a few dominant components representing the MMR with a high signal-to-noise ratio and similar topography to that of the sensor space analysis using the subtraction approach. In contrast, infomax or weighted-infomax revealed many minor or pseudo components as constituents of the MMR. Our new approach may assist in using the MMR in basic and clinical research.Clinical Relevance-Our proposed method opens a new and potentially useful way to analyze event-related MEG/EEG data.

A novel method for extracting interictal epileptiform discharges in multi-channel MEG: Use of fractional type of blind source separation.

OBJECTIVE: Visual inspection of interictal epileptiform discharges (IEDs) in multi-channel MEG requires a time-consuming evaluation process and often leads to inconsistent results due to variability of IED waveforms. Here, we propose a novel extraction method for IEDs using a T/k type of blind source separation (BSST/k). METHODS: We applied BSST/k with seven patients with focal epilepsy to test the accuracy of identification of IEDs. We conducted comparisons of the results of BSS components with those obtained by visual inspection in sensor-space analysis. RESULTS: BSST/k provided better signal estimation of IEDs compared with sensor-space analysis. Importantly, BSST/k was able to uncover IEDs that could not be detected by visual inspection. Furthermore, IED components were clearly extracted while preserving spike and wave morphology. Variable IED waveforms were decomposed into one dominant component. CONCLUSIONS: BSST/k was able to visualize the spreading signals over multiple channels into a single component from a single epileptogenic zone. BSST/k can be applied to focal epilepsy with a simple parameter setting. SIGNIFICANCE: Our novel method was able to highlight IEDs with increased accuracy for identification of IEDs from multi-channel MEG data.

Neurodynamics of somatosensory cortices studied by magnetoencephelography.

From the viewpoint of statistical inverse problems, identification of transfer functions in feedback models is applied for neurodynamics of somatosensory cortices, and brain communication among active regions can be expressed in terms of transfer functions. However, brain activities have been investigated mainly by averaged waveforms in the conventional magnetoencephalography analysis, and thus brain communication among active regions has not yet been identified. It is shown that brain communication among two more than three brain regions is determined, when fluctuations related to concatenate averaged waveforms can be obtained by using a suitable blind source separation method. In blind identification of feedback model, some transfer functions or their impulse responses between output variables of current dipoles corresponding to active regions are identified from reconstructed time series data of fluctuations by the method of inverse problem. Neurodynamics of somatosensory cortices in 5 Hz median nerve stimuli can be shown by cerebral communication among active regions of somatosensory cortices in terms of impulse responses of feedback model.

Evoked magnetic fields of magnetoencephalography and their statistical property.

In an evoked magnetic field of magnetoencephalography a wave form is calculated by averaging. We propose that the wave form is deterministic in the case of 5 Hz periodical stimuli. We have found with statistical accuracy that the wave form of a somatosensory evoked magnetic field is deterministic in 5 Hz periodical median nerve stimuli, since any stationary process is decomposed into a deterministic part and a nondeterministic part from the Wold decomposition theorem. For the decorrelation method of blind source separation we have obtained several components which have nonzero wave forms. Via the selected components time series data of a somatosensory evoked magnetic field generated from somatosensory cortexes have been separated from background brain noise by using a T/k (fractional) type decorrelation method.

Dynamical activities of primary somatosensory cortices studied by magnetoencephalography.

A blind identification method of transfer functions in feedback systems is introduced for examination of dynamical activities of cortices by magnetoencephalography study. Somatosensory activities are examined in 5 Hz periodical median nerve stimulus. In the present paper, we will try two careful preprocessing procedures for the identification method to obtain impulse responses between primary somatosensory cortices. Time series data of the somatosensory evoked field are obtained by using a blind source separation of the T/k type (fractional) decorrelation method. Time series data of current dipoles of primary somatosensory cortices are transformed from the time series data of the somatosensory evoked field by the inverse problem. Fluctuations of current dipoles of them are obtained after elimination of deterministic periodical evoked waveforms. An identification method based on feedback system theory is used for estimation of transfer functions in a feedback model from obtained fluctuations of currents dipoles of primary somatosensory cortices. Dynamical activities between them are presented by Bode diagrams of transfer functions and their impulse responses: the time delay of about 30 ms via corpus callosum is found in the impulse response of identified transfer function.