Choice of Magnetometers and Gradiometers after Signal Space Separation

Garcés, Pilar; López‐Sanz, David; Maestú, Fernando; Pereda, Ernesto

doi:10.3390/s17122926

Cited by 96 publications

(78 citation statements)

References 38 publications

(52 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…SSS decomposes the MEG signal into extracranial and intracranial sources and renders the data rank-deficient. Once applied, magnetometers and gradiometers become linear combinations of approximately 70 common SSS components, hence, become interchangeable (Garcés et al, 2017). For simplicity, we conducted all analyses on magnetometers.…”

Section: Mne Model For Regression With Source Localizationmentioning

confidence: 99%

“…When performing SSS, one actually combines the information of independent gradiometer and magnetometer sensor arrays into one latent space of roughly 65 dimensions that is roughly 20% of the dimensionality of both sensor arrays (306 sensors in total). Even when analyzing the magnetometers only after SSS, one will also access the extra information from the gradiometers (Garcés et al, 2017). SSP on the other hand is less invasive and is applied separately to magnetometers and gradiometers.…”

Section: How Important Is Preprocessing For Cross-subject Prediction?mentioning

confidence: 99%

See 1 more Smart Citation

Predictive regression modeling with MEG/EEG: from source power to signals and cognitive states

Sabbagh

Ablin

Varoquaux

et al. 2019

Preprint

View full text Add to dashboard Cite

Predicting biomedical outcomes from Magnetoencephalography and Electroencephalography (M/EEG) is central to applications like decoding, brain-computer-interfaces (BCI) or biomarker development and is facilitated by supervised machine learning. Yet most of the literature is concerned with within-subject classification. Here, we focus on predicting continuous outcomes from M/EEG signal power across subjects. Considering different generative mechanisms for M/EEG signals and the biomedical outcome, we propose statistically-consistent predictive models that avoid sourcereconstruction based on the covariance as representation. Our mathematical analysis and ground truth simulations demonstrated that consistent parameter estimation can be obtained with Source Power Comodulation (SPoC) supervised spatial filtering or by embedding with Riemannian geometry. Additional simulations revealed that Riemannian methods were more robust to model violations, in particular geometric distortions induced by individual anatomy. To estimate the relative contribution of brain dynamics and anatomy to prediction performance, we propose a novel model inspection procedure based on biophysical forward modeling. Applied to cross-subject prediction, the analysis revealed that the Riemannian model better exploited anatomical information while sensitivity to brain dynamics was similar across methods. We then probed the robustness of the models across different data cleaning options. Environmental denoising was globally important but Riemannian models were strikingly robust and continued performing even without preprocessing. Our results suggest each method has its niche: SPoC is practical for within-subject prediction while the Riemannian model may enable simple end-to-end learning.

show abstract

Section: Mne Model For Regression With Source Localizationmentioning

confidence: 99%

Section: How Important Is Preprocessing For Cross-subject Prediction?mentioning

confidence: 99%

Predictive regression modeling with MEG/EEG: from source power to signals and cognitive states

Sabbagh

Ablin

Varoquaux

et al. 2019

Preprint

View full text Add to dashboard Cite

show abstract

“…to attenuate external noise from the MEG signal (mainly 16.6Hz, and 50Hz plus harmonics) and 444 realign data to a common standard head position ("-trans default" Maxfilter parameter) across 445 different blocks based on the measured head position at the beginning of each block 45 . The rest of 446 the subsequent analysis was performed on magnetometers only, given the mixing of information 447 between the two sensors types after the Maxfilter step 46 . 448 Data analysis was carried out with scripts written in-house, using the Fieldtrip toolbox 47 449 (git version 20170919).…”

Section: Supplementary Materialsmentioning

confidence: 99%

Phantom auditory perception (tinnitus) is characterised by stronger anticipatory auditory predictions

Partyka

Demarchi

Roesch

et al. 2019

Preprint

View full text Add to dashboard Cite

19How phantom perceptions arise and the factors that make individuals prone to such 20 experiences are not well understood. An attractive phenomenon to study these questions is tinnitus, 21 a very common auditory phantom perception which is not explained by hyperactivity in the 22 auditory pathway alone. Our framework posits that a predisposition to developing (chronic) 23 tinnitus is dependent on individual traits relating to the formation and utilization of sensory 24 predictions. Predictions of auditory stimulus frequency (remote from tinnitus frequency) were 25 studied using a paradigm parametrically modulating regularity (i.e. predictability) of tone 26 sequences and applying decoding techniques on magnetoencephalographic (MEG) data. For 27 processes likely linked to short-term memory, individuals with tinnitus showed an enhanced 28 anticipatory prediction pattern associated with increasing sequence regularity. In contrast, 29 individuals without tinnitus engaged the same processes following the onset of the to-be-decoded 30 sound. We posit that this tendency to optimally anticipate static and changing auditory inputs may 31 determine which individuals faced with persistent auditory pathway hyperactivity factor it into 32 auditory predictions, and thus perceive it as tinnitus. While our study constitutes a first step relating 33 vulnerability to tinnitus with predictive processing, longitudinal studies are needed to confirm the 34 predisposition model of tinnitus development. 35 36 37 38 39Research along these lines has focused mostly on probable "neural correlate" candidates 54 of tinnitus such as increased spontaneous firing rate or enhanced neural synchrony. The altered 55 gain model of tinnitus is substantially supported by studies in animals 11 , despite the obvious 56 challenges in obtaining subjective reports. In humans the supporting evidence for this model is 57 less apparent, partly because (contrary to animal models) the research is focused on chronic rather 58 than acute tinnitus, but also due to a lack of understanding as to how measures commonly obtained 59 in humans (such as oscillatory power in M/EEG or BOLD in fMRI) can be translated to those used 60 to support the altered gain model. Based on human and animal works in other domains 12 , reduced 61 ongoing alpha or increased gamma in auditory regions pertinent for phantom sounds (for other 62 auditory phantom percepts see 13,14 ) may be relevant to perception of tinnitus. However, the 63 empirical evidence is inconclusive 15,16 . With the exception of technical or practical issues that 64 may complicate a convincing confirmation of the altered gain model in humans, other observations 65 speak in favour of its explanatory insufficiency 17 : 1) Only a fraction of individuals who suffer a 66 hearing impairment will experience tinnitus (~70% following sensorineural hearing loss; see 18-67 20 ). 2) The onset of tinnitus and the onset of the hearing loss often occur at different times. 3) Not 68 all cases of acute tinnitus will become chronic....

show abstract

“…Gradiometers sample from a smaller area than magnetometers, which is important for ensuring a separability of nodes as expected by network models [48]. Furthermore, gradiometers are typically less susceptible to noise than magnetometers [49]. We combined data from planar gradiometers in the voltage domain using the 'sum' method from Fieldtrip's ft combine planar() function (http://www.fieldtriptoolbox.org/).…”

Section: Preprocessingmentioning

confidence: 99%

Learning in brain-computer interface control evidenced by joint decomposition of brain and behavior

Stiso¹,

Corsi²,

Vettel³

et al. 2019

Preprint

View full text Add to dashboard Cite

Motor imagery-based brain-computer interfaces (BCIs) use an individuals ability to volitionally modulate localized brain activity, often as a therapy for motor dysfunction or to probe causal relations between brain activity and behavior. However, many individuals cannot learn to successfully modulate their brain activity, greatly limiting the efficacy of BCI for therapy and for basic scientific inquiry. Formal experiments designed to probe the nature of BCI learning have offered initial evidence that coherent activity across spatially distributed and functionally diverse cognitive systems is a hallmark of individuals who can successfully learn to control the BCI. However, little is known about how these distributed networks interact through time to support learning. Here, we address this gap in knowledge by constructing and applying a multimodal network approach to decipher brain-behavior relations in motor imagery-based brain-computer interface learning using magnetoencephalography. Specifically, we employ a minimally constrained matrix decomposition method -non-negative matrix factorization -to simultaneously identify regularized, covarying subgraphs of functional connectivity, to assess their similarity to task performance, and to detect their time-varying expression. We find that learning is marked by diffuse brain-behavior relations: good learners displayed many subgraphs whose temporal expression tracked performance. Individuals also displayed marked variation in the spatial properties of subgraphs such as the connectivity between the frontal lobe and the rest of the brain, and in the temporal properties of subgraphs such as the stage of learning at which they reached maximum expression. From these observations, we posit a conceptual model in which certain subgraphs support learning by modulating brain activity in regions important for sustaining attention. To test this model, we use tools that stipulate regional dynamics on a networked system (network control theory), and find that good learners display a single subgraph whose temporal expression tracked performance and whose architecture supports easy modulation of brain regions important for attention. The nature of our contribution to the neuroscience of BCI learning is therefore both computational and theoretical; we first use a minimally-constrained, individual specific method of identifying mesoscale structure in dynamic brain activity to show how global connectivity and interactions between distributed networks supports BCI learning, and then we use a formal network model of control to lend theoretical support to the hypothesis that these identified subgraphs are well suited to modulate attention.

show abstract

Choice of Magnetometers and Gradiometers after Signal Space Separation

Cited by 96 publications

References 38 publications

Predictive regression modeling with MEG/EEG: from source power to signals and cognitive states

Predictive regression modeling with MEG/EEG: from source power to signals and cognitive states

Phantom auditory perception (tinnitus) is characterised by stronger anticipatory auditory predictions

Learning in brain-computer interface control evidenced by joint decomposition of brain and behavior

Contact Info

Product

Resources

About