Measuring MEG closer to the brain: Performance of on-scalp sensor arrays

“FreeSurfer software (Dale et al., 1999; Fischl et al., 1999a; Fischl, 2012) was used to segment the cortical mantles from individual MRIs.”

“FreeSurfer software (Dale et al., 1999; Fischl et al., 1999a; Fischl, 2012) was used to segment the cortical mantles from individual MRIs.”

“FreeSurfer software (Dale et al., 1999; Fischl et al., 1999a; Fischl, 2012) was used to segment the cortical mantles from individual MRIs.”

“2.3 Minimum-norm estimate Minimum-l2-norm estimator (Dale and Sereno, 1993; Hämäläinen and Ilmoniemi, 1994) can be used to estimate underlying sources from measured data.”

“For each subject, we used the watershed algorithm (Ségonne et al., 2004) implemented in FreeSurfer and MNE software (Gramfort et al., 2014) to segment the brain, skull and scalp compartments.”

“, 2004) implemented in FreeSurfer and MNE software (Gramfort et al., 2014) to segment the brain, skull and scalp compartments.”

“2.3 Minimum-norm estimate Minimum-l2-norm estimator (Dale and Sereno, 1993; Hämäläinen and Ilmoniemi, 1994) can be used to estimate underlying sources from measured data.”

“2.3 Minimum-norm estimate Minimum-l2-norm estimator (Dale and Sereno, 1993; Hämäläinen and Ilmoniemi, 1994) can be used to estimate underlying sources from measured data.”

“Alternative sensors such as optically-pumped, or atomic, magnetometers (OPMs) (Budker and Kimball, 2013; Budker and Romalis, 2007) and high-Tc SQUIDs (Öisjöen et al., 2012) have recently reached sensitivities required for detecting neuromagnetic fields.”

“Particularly OPMs operating in the spin-exchange relaxation-free (SERF) regime (Allred et al., 2002) could become a feasible, noncryogenic alternative to SQUIDs for neuromagnetic measurements as the combined dimensions and sensitivies of the SERF-OPMs are approaching those of SQUIDs (∼4–6 cm(2) and ∼3 − 5 fT/ Hz ).”

“We also compared the topographies obtained with a linear collocation (LC) (De Munck, 1992) and LG BEM (Mosher et al., 1999) across the subjects.”

“2.1 Physics When bioelectromagnetic fields are modeled in the macroscopic scale, the time-dependent terms in Maxwell’s equations can be omitted (Plonsey and Heppner, 1967).”

“The authors computed the magnetic field with local-spheres head model (Huang et al., 1999) in a single head geometry.”

“Thus, the field computation for normal-component-measuring OPMs could potentially be done with a LC BEM, which is computationally lighter than the more accurate solvers such as LG BEM or symmetric BEM (Kybic et al., 2005).”

“Thus, compared to the state-of-the-art OPMs, our hypothetical OPMs are optimistic in design; however, we anticipate that the distance between the vapor cell and the outer surface of the sensor will shrink in the next generation OPMs due to advances in sensor packaging and in the utilization of microfabrication (Liew e…”

“To evaluate the differences between the arrays in linear distributedsource estimation, we used metrics derived from the resolution matrix (Hauk et al., 2011; Liu et al., 2002; Molins et al., 2008; de Peralta Menendez et al., 1997; Stenroos and Hauk, 2013).”

“We assumed a noise level of 6 fT/ for the OPM sensor based on the work by Shah and Wakai (2013) who reported such a field resolution for their sensor in optimal conditions.”

“When the conducting region is modeled to consist of piecewise homogeneous regions, the magnetic field is given by the Geselowitz formula; (Geselowitz, 1970) where Bp(r) and Bv(r) are the magnetic fields due to primary and volume currents, μ0 is the permeability of free space, and , and are the…”

“The effect of common volume-conductor-model errors and simplifications on both nOPM and tOPM could be assessed by comparing geometrically correct and erroneous simplified models to more detailed reference models (Stenroos et al., 2014; Stenroos and Nummenmaa, 2016; Vorwerk et al., 2014).”

“The regularization parameter has been suggested to be chosen as (Lin et al., 2006a,b) where Nc is the number of sensors, SNR is the estimated signal-to-noise ratio, tr(·) is trace of a matrix and L is the whitened lead-field matrix: .”

“For each subject, we used the watershed algorithm (Ségonne et al., 2004) implemented in FreeSurfer and MNE software (Gramfort et al., 2014) to segment the brain, skull and scalp compartments.”

“We also compared the topographies obtained with a linear collocation (LC) (De Munck, 1992) and LG BEM (Mosher et al., 1999) across the subjects.”

“To evaluate the differences between the arrays in linear distributed-source estimation, we used metrics derived from the resolution matrix (Hauk et al., 2011; Liu et al., 2002; Molins et al., 2008; de Peralta Menendez et al., 1997; Stenroos and Hauk, 2013).”

“To evaluate the differences between the arrays in linear distributed-source estimation, we used metrics derived from the resolution matrix (Hauk et al., 2011; Liu et al., 2002; Molins et al., 2008; de Peralta Menendez et al., 1997; Stenroos and Hauk, 2013).”

“In addition, we investigated the overlap of the sensor lead fields as a function of the sensor distance (de Munck et al., 1992): for each sensor in the array, we calculated the distances and lead-field correlations between that sensor and all other sensors in the array.”

“For each subject, the lead-field matrix L was computed using a linear Galerkin BEM formulated with Isolated Source Approach (ISA) (Stenroos and Sarvas, 2012).”

“The aim of this study was to assess the possible benefits of on-scalp MEG arrays that can be constructed using novel magnetic field sensors, e.g., optically-pumped magnetometers (OPMs) or high-Tc SQUIDs.”

“The effect of common volume-conductor-model errors and simplifications on both nOPM and tOPM could be assessed by comparing geometrically correct and erroneous simplified models to more detailed reference models (Stenroos et al., 2014; Stenroos and Nummenmaa, 2016; Vorwerk et al., 2014).”

“To evaluate the differences between the arrays in linear distributed-source estimation, we used metrics derived from the resolution matrix (Hauk et al., 2011; Liu et al., 2002; Molins et al., 2008; de Peralta Menendez et al., 1997; Stenroos and Hauk, 2013).”

“Previously, the performance of various MEG arrays has been assessed by computing the total information Itot conveyed by the array (Kemppainen and Ilmoniemi, 1989; Nenonen et al., 2004; Schneiderman, 2014).”

“To complete forward model -based metrics and quantify the performance of the array with a single number, we calculated the total information conveyed by the array (Kemppainen and Ilmoniemi, 1989).”

“Similar observations have been made for cardiomagnetic fields (Arturi et al., 2004).”

“…indicate that the normal and tangential components carry independent information as the information capacity of the aOPM is higher than the information capacities of its sub-arrays nOPM and tOPM, indicating that these measurements are not redundant; see also de Munck and Daffertshofer (2012).”

“Simulations [12,13] show that this effect provides a four- to fivefold signal enhancement in many cortical areas.”

“…advances in sensors, additional neuroscientific motivation for OPMs came from simulation studies examining issues such as improvements in SNR, spatial resolution and the requirements necessary for accurate source reconstruction (Boto et al., 2016a; Iivanainen et al., 2017a; Zetter et al., 2017).”