Inverse Source Modeling Research Articles

Mapping InSAR deformation of low and moderate earthquakes

Abstract Low-to-moderate magnitude earthquakes often induce small ground displacement. For such events, ground deformation fields detected by SAR interferometry can be masked or not clearly discernible from the fringes distribution because of the presence of error sources, such as atmospheric artifacts and topographic residuals. We show two examples of low-moderate magnitude earthquakes, for which we adopted a new automatic tool for calculating the fringe pattern stemming from seismological data. The tool estimates the extent and the geographic position of the deformation by running a forward model of the seismic source, thus identifying the best SAR frame to be collected and the expected surface effects to better figure-out the outcomes of InSAR processing. We present the Mw 5.7 occurred in Greece and the Mw 5.4 occurred in Zagreb on 21 and 22 March 2020, respectively. For the Greek earthquake, the tool predicted a deformation close to the InSAR product, and gave evidences of atmospheric disturbances, thus providing information for inverse source modelling. The tool in the Zagreb event was used to infer the extent and location of the ground motion, that were used to identify the best SAR pair to be processed, being the SAR frames edge very close to the epicenter.

Seizure localization using EEG analytical signals

ObjectiveLocalization of epileptic seizures, usually characterized by abnormal hypersynchronous wave patterns from the cortex, remains elusive. We present a novel, robust method for automatic localization of seizures on the scalp from clinical electroencephalogram (EEG) data. MethodsSeizure patient EEG data was decomposed via the Hilbert Transform and processed through the following methodology: sorting the analytic amplitude (AA) in the time instance, locating the maximum amplitude within the vector of channels, cross-correlating amplitude values in the time index with the channel vector. The channel with highest AA value in time was located. ResultsOur approach provides an automated way to isolate the epi-genesis of seizure events with 93.3% precision and 100% sensitivity. The method differentiates seizure-related neural activity from other common EEG noise artifacts (e.g., blinks, myogenic noise). ConclusionsWe evaluated performance characteristics of our source location methodology utilizing both phase and energy of EEG signals from patients who exhibited seizure events. Feasibility of the new algorithm is demonstrated and confirmed. SignificanceThe proposed method contributes to high-performance scalp localization for seizure events that is more straightforward and less computationally intensive than other methods (e.g., inverse source modeling). Ultimately, it may aid clinicians in providing improved patient diagnosis.

Electromagnetic Machine Learning for Inverse Modeling Using the Spatial Singularity Expansion Method

In this article, we propose a general technique that utilizes a unified electromagnetic and machine learning (ML) technique for inverse modeling and antenna characterization. The recently developed spatial singularity expansion method (S-SEM) is deployed to explicate the electromagnetic behavior of antennas in the form of an accurate digital signal processing (DSP) model. An ML framework is devised and combined with the S-SEM-based DSP model to design a novel inverse source modeling algorithm. The combined S-SEM-ML system departs from the state-of-the-art approaches in being capable of processing far-field data in order to jointly estimate the surface current distribution on the examined radiators, in addition to reconstructing their geometrical details. Various straight and bent wire systems are investigated, including single and multiple array configurations. A study on the impact of additive free-space noise on the estimation process is also presented. Finally, a measurement apparatus is provided to validate the proposed S-SEM-ML system.

T15. Imaging fast neuronal depolarization in 3D during seizures with electrical impedance tomography and scalp and intracranial depth electrodes

Electrical Impedance Tomography (EIT) is an imaging technique where internal conductivity changes in an object are reconstructed from measurements with arrays of external electrodes. It has the potential to improve epileptogenic zone localisation as seizure activity changes local impedance in two ways. ‘Fast’ changes, ∼1% over millseconds, are due to the opening of ion channels during synchronised neuronal depolarisation, and ‘slow’ changes, up to 10% over seconds, are due to cell swelling. During SEPs in anaethetised rats, EIT could image evoked activity in S1 with a resolution of 1 ms and <200 μm using 30 electrode epicortical arrays. EIT could offer improved resolution over the current technique of inspection of potentials with depth electrodes (SEEG) in patients undergoing presurgical seizure evaluation. EIT has the advantage that it is not sensitive to dipole orientation and coverage within a volume enclosed by electrodes is much better than for potential recording. The purpose of this study was to undertake a computer modelling study to compare fast neural EIT with SEEG and EEG inverse source modelling in subjects with epilepsy and intracranial electrodes. The location and shape accuracy of reconstructed changes were assessed in 3 subjects with 48–72 depth electrode contacts and 32 scalp electrodes with EIT (1.7 kHz, 50 μA intracranial, 250 μA scalp), EEG inverse source modelling (ISM), and spike detection on depth electrodes (SEEG), using FEM meshes of ∼9 M tetrahedral elements generated from CT-MRI. EIT and inverse source reconstructions were performed on ∼30,000 hexahedral element meshes with zeroth order Tikhonov regularisation and noise-based image weighting. Seizure onset was simulated by 5 mm radius realistic perturbations of 1% and 10% or dipoles which produced 2 mV when 5 mm away from an electrode (5 within and 5 outside the volume enclosed by depth electrodes in each subject). EIT produced the best localisation. It was far superior to ISM and also provided accurate imaging for dipole sources not apparent on SEEG because of distance from the nearest electrode or tangential orientation. For perturbations placed within the depth electrodes, localisation accuracy was 5.2 ± 1.8 mm (1% change) <5 mm (10% change) for EIT and 46.2 ± 25.8 mm for ISM with radially oriented sources. For SEEG, 7/15 radially oriented sources did not reach a visualisable threshold set as 250 μV which occurred when they were >11 mm distant from the nearest electrode. For the sources in the opposite hemisphere, it was 29.6 ± 38.7 mm (1% EIT), 26.1 ± 36.2 (10% EIT), 54.0 ± 26.2 mm (ISM) and 0/15(SEEG). In this modelling study, fast and slow neural EIT with depth electrode offers a new method for localisation of seizure onset and propagation. It employs tiny injected currents which do not damage the brain or alter cerebral function. It offers a potentially valuable additional method for presurgical epilepsy evaluation. A human clinical trial is in progress.

Feasibility of imaging epileptic seizure onset with EIT and depth electrodes

Imaging ictal and interictal activity with Electrical Impedance Tomography (EIT) using intracranial electrode mats has been demonstrated in animal models of epilepsy. In human epilepsy subjects undergoing presurgical evaluation, depth electrodes are often preferred. The purpose of this work was to evaluate the feasibility of using EIT to localise epileptogenic areas with intracranial electrodes in humans. The accuracy of localisation of the ictal onset zone was evaluated in computer simulations using 9M element FEM models derived from three subjects. 5 mm radius perturbations imitating a single seizure onset event were placed in several locations forming two groups: under depth electrode coverage and in the contralateral hemisphere. Simulations were made for impedance changes of 1% expected for neuronal depolarisation over milliseconds and 10% for cell swelling over seconds. Reconstructions were compared with EEG source modelling for a radially orientated dipole with respect to the closest EEG recording contact. The best accuracy of EIT was obtained using all depth and 32 scalp electrodes, greater than the equivalent accuracy with EEG inverse source modelling. The localisation error was 5.2 ± 1.8, 4.3 ± 0 and 46.2 ± 25.8 mm for perturbations within the volume enclosed by depth electrodes and 29.6 ± 38.7, 26.1 ± 36.2, 54.0 ± 26.2 mm for those without (EIT 1%, 10% change, EEG source modelling, n = 15 in 3 subjects, p < 0.01). As EIT was insensitive to source dipole orientation, all 15 perturbations within the volume enclosed by depth electrodes were localised, whereas the standard clinical method of visual inspection of EEG voltages, only localised 8 out of 15 cases. This suggests that adding EIT to SEEG measurements could be beneficial in localising the onset of seizures.

PB 11 Phase-coupling optimization (PCO) – a new algorithm for the analysis and localization of phase-coupling in multivariate data

Introduction Phase-coupling between neuronal oscillations in the brain has been hypothesized as a mechanism for interactions between brain areas; moreover, phases of neuronal oscillations were found related to performance in memory and perception tasks. However, in multivariate data, such as multi-channel EEG or MEG, the analysis of phase-coupling remains challenging. Sensor-space analysis is suboptimal regarding the signal-to-noise ratio (SNR) and might fail to localize the sources of phase-coupling appropriately. Here, we introduce phase-coupling optimization (PCO), an algorithm seeking spatial filters that maximize the coupling of oscillatory phases to an independent variable (e.g., detected vs. undetected events in a perception task). The resulting spatial filters/patterns can then be used for inverse source modeling. Methods Due to its simplicity, the “mean vector length” measure was chosen to quantify the degree of phase-coupling and is being maximized using the quasi-Newton Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithm. To avoid local minima, the optimization is restarted several times from low-discrepancy pseudo-random initial starting points. As regularization, spatio-spectral decomposition (SSD) reduces the search-space while maximizing the SNR in the selected frequency band. The performance of the algorithm was verified using realistic forward-model EEG simulations. Results Simulations demonstrated that PCO, together with SSD as pre-processing, can reliably localize the sources of phase-coupling in multi-channel EEG and characterize the coupling relation down to a SNR of −10 dB (i.e., a power ratio of 0.1) using 500 simulated task repetitions. Additionally, PCO is superior to a multiple regression approach and by a large margin outperforms sensor-space analysis based on a current source density estimation. Discussion The analysis of phase-coupling in sensor space is suboptimal due to the SNR of the signal and might even lead to erroneous source localization. Techniques such as beamforming on the other hand, require a well-founded a priori assumption on the spatial origin of phase-coupling and complex inverse problem solvers. PCO seeks the optimal spatial projection for the analysis of phase-coupling and demonstrated its reliability even for very low SNR. Hence, we regard it as a promising tool for the analysis of neuronal oscillations. Significance PCO increases the sensitivity and reliability of phase-coupling analyses. The obtained spatial patterns can be used for inverse source modeling.

"Cerebellar contribution to visuo-attentional alpha rhythm: insights from weightlessness".

Human brain adaptation in weightlessness follows the necessity to reshape the dynamic integration of the neural information acquired in the new environment. This basic aspect was here studied by the electroencephalogram (EEG) dynamics where oscillatory modulations were measured during a visuo-attentional state preceding a visuo-motor docking task. Astronauts in microgravity conducted the experiment in free-floating aboard the International Space Station, before the space flight and afterwards. We observed stronger power decrease (~ERD: event related desynchronization) of the ~10 Hz oscillation from the occipital-parietal (alpha ERD) to the central areas (mu ERD). Inverse source modelling of the stronger alpha ERD revealed a shift from the posterior cingulate cortex (BA31, from the default mode network) on Earth to the precentral cortex (BA4, primary motor cortex) in weightlessness. We also observed significant contribution of the vestibular network (BA40, BA32, and BA39) and cerebellum (lobule V, VI). We suggest that due to the high demands for the continuous readjustment of an appropriate body posture in free-floating, this visuo-attentional state required more contribution from the motor cortex. The cerebellum and the vestibular network involvement in weightlessness might support the correction signals processing necessary for postural stabilization, and the increased demand to integrate incongruent vestibular information.

Hypervigilance-avoidance in children with anxiety disorders: magnetoencephalographic evidence.

An altered pattern of threat processing is deemed critical for the development of anxiety disorders (AD). According to the hypervigilance-avoidance hypothesis, AD patients show hypervigilance to threat cues at early stages of processing but avoid threat cues at later stages of processing. Consistently, adults with AD show enhanced neurophysiological responses to threat in early time windows and reduced responses to threat in late time windows. The presence of such a hypervigilance-avoidance effect and its underlying neural sources remain to be determined in clinically anxious children. Twenty-three children diagnosed with an AD and 23 healthy control children aged 8-14years saw faces with angry and neutral expressions while whole-head magnetoencephalography (MEG) was recorded. Neural sources were estimated based on L2-Minimum Norm inverse source modeling and analyzed in early, midlatency, and late time windows. In visual cortical regions, early threat processing was relatively enhanced in patients compared to controls, whereas this relation was inverted in a late interval. Consistent with the idea of affective regulation, the right dorsolateral prefrontal cortex revealed relatively reduced inhibition of early threat processing but revealed enhanced inhibition at a late interval in patients. Both visual-sensory and prefrontal effects were correlated with individual trait anxiety. These results support the hypothesis of early sensory hypervigilance followed by later avoidance of threat in anxiety disordered children, presumably modulated by early reduced and later enhanced prefrontal inhibition. This neuronal hypervigilance-avoidance pattern unfoldsgradually with increasing trait anxiety, reflecting a progressively biased allocation of attention to threat.

Walking reduces sensorimotor network connectivity compared to standing

BackgroundConsiderable effort has been devoted to mapping the functional and effective connectivity of the human brain, but these efforts have largely been limited to tasks involving stationary subjects. Recent advances with high-density electroencephalography (EEG) and Independent Components Analysis (ICA) have enabled study of electrocortical activity during human locomotion. The goal of this work was to measure the effective connectivity of cortical activity during human standing and walking.MethodsWe recorded 248-channels of EEG as eight young healthy subjects stood and walked on a treadmill both while performing a visual oddball discrimination task and not performing the task. ICA parsed underlying electrocortical, electromyographic, and artifact sources from the EEG signals. Inverse source modeling methods and clustering algorithms localized posterior, anterior, prefrontal, left sensorimotor, and right sensorimotor clusters of electrocortical sources across subjects. We applied a directional measure of connectivity, conditional Granger causality, to determine the effective connectivity between electrocortical sources.ResultsConnections involving sensorimotor clusters were weaker for walking than standing regardless of whether the subject was performing the simultaneous cognitive task or not. This finding supports the idea that cortical involvement during standing is greater than during walking, possibly because spinal neural networks play a greater role in locomotor control than standing control. Conversely, effective connectivity involving non-sensorimotor areas was stronger for walking than standing when subjects were engaged in the simultaneous cognitive task.ConclusionsOur results suggest that standing results in greater functional connectivity between sensorimotor cortical areas than walking does. Greater cognitive attention to standing posture than to walking control could be one interpretation of that finding. These techniques could be applied to clinical populations during gait to better investigate neural substrates involved in mobility disorders.

The neural basis of cognitive change: Reappraisal of emotional faces modulates neural source activity in a frontoparietal attention network

Emotions can be regulated effectively via cognitive change, as evidenced by cognitive behavioural therapy. The neural correlates of cognitive change were investigated using reappraisal, a strategy that involves the reinterpretation of emotional stimuli. Hemodynamic studies revealed cortical structures involved in reappraisal and highlighted the role of the prefrontal cortex in regulating subcortical affective processing. Studies using event-related potentials elucidated the timing of reappraisal by showing effective modulation of the Late Positive Potential (LPP) after 300ms but also even earlier effects. The present study investigated the spatiotemporal dynamics of the cortical network underlying cognitive change via inverse source modelling based on whole-head magnetoencephalography (MEG). During MEG recording, 28 healthy participants saw angry and neutral faces and followed instructions designed to down- or up-regulate emotions via reappraisal. Differences between angry and neutral face processing were specifically enhanced during up-regulation, first in the parietal cortex during M170 and across the whole cortex during LPP-M, with particular involvement of the parietal and dorsal prefrontal cortex regions. Thus, our data suggest that the reappraisal of emotional faces involves specific modulations in a frontoparietal attention network.

Corticomuscular coherence revealed during treadmill walking: further evidence of supraspinal control in human locomotion

Corticomuscular coherence revealed during treadmill walking: further evidence of supraspinal control in human locomotion

Can principal component analysis reliably identify the temporal response of cortical area V1?

The early CI component and its polarity inversion as a stimulus moves from inferior to superior visual field is the basis for many evoked response potential (ERP) based claims of attention and other perceptual effects involving V1 activation. However, recent results have undermined the validity those criteria. Ales et al. (2010) have shown "polarity inversion is an unreliable diagnostic for activity originating in striate cortex". The fact that the CI is the earliest response peak is also not indicative of a V1 signal since recent ERP source modeling demonstrate that the signals from V1 and V2 begin at about the same time (Ales et al, 2009). With historical ERP measures of V1 activation discredited, does any method offer reliable isolation of the striate cortex response besides inverse source modeling along with fMRI to constrain the solution space? Zhang and Hood (2004) suggested principal component analysis provides a means of deriving the V1 response when combined with multi-focal stimulation. The change in the first component signal strength with stimulus position was consistent with the cruciform model of V1. However, we now know the cruciform model is flawed, so we revisited the utility of using PCA to isolate V1. In two subjects, using high density multi-focal stimuli with a 96 electrode recording array we compared the V1 temporal response based on PCA verses inverse source modeling methods. The concordance between the V1 response derived from PCA and fMRI constrained inverse source modeling was high. However, in 21 subjects, simulated ERPs using boundary element forward models and fMRI based source positions corresponding to a 96 patch multi-focal stimulus demonstrated that the first PCA component of simulated ERP's do not always match the V1 source time functions. In conclusion, PCA of the multi-focal ERP does not provide a reliable estimate of an individual's V1 response. Meeting abstract presented at VSS 2012

A comparison of the accuracy of EEG head nets in inverse source modelling of the EEG using a realistic head phantom

A comparison of the accuracy of EEG head nets in inverse source modelling of the EEG using a realistic head phantom

Right hemisphere control of visuospatial attention: line-bisection judgments evaluated with high-density electrical mapping and source analysis☆

The “line-bisection” task has proven an especially useful clinical tool for assessment of spatial neglect syndrome in neurological patients. Here, we investigated the neural processes involved in performing this task by recording high-density event-related potentials from 128 scalp electrodes in normal observers. We characterized a robust net negative potential from 170–400 ms poststimulus presentation that correlates with line-bisection judgments. Topographic mapping shows three distinct phases to this negativity. The first phase (∼170–190 ms) has a scalp distribution exclusively over the right parieto-occipital and lateral occipital scalp, consistent with generators in the region of the right temporo-parietal junction and right lateral occipital cortices. The second phase (∼190–240 ms) sees the emergence of a second negative focus over the right central parietal scalp, consistent with subsequent involvement of right superior parietal cortices. In the third phase (∼240–400 ms), the topography becomes dominated by this right central parietal negativity. Inverse source modeling confirmed that right hemisphere lateral occipital, inferior parietal, and superior parietal regions were the likeliest generators of the bulk of the activity associated with this effect. The line stimuli were also presented at three contrast levels (3, 25, and 100%) in order to manipulate both the latency of stimulus processing and the relative contributions from magnocellular and parvocellular inputs. Through this manipulation, we show that the line-bisection effect systematically tracks/follows the latency of the N1 component, which is considered a temporal marker for object processing in the ventral visual stream. This pattern of effects suggests that this task invokes an allocentric (object-based) form of visuospatial attention. Further, at 3% contrast, the line-bisection effect was equivalent to the effects seen at higher contrast levels, suggesting that parvocellular inputs are not necessary for successful performance of this task.