A Gaussian process model of human electrocorticographic data

Abstract

Human Super EEG 1 entails measuring ongoing activity from every cell in a living human brain at millisecond-scale temporal resolutions. Although direct cell-by-cell Super EEG recordings are impossible using existing methods, here we present a technique for inferring neural activity at arbitrarily high spatial resolutions using human intracranial electrophysiological recordings. Our approach, based on Gaussian process regression, relies on two assumptions. First, we assume that some of the correlational structure of people's brain activity is similar across individuals. Second, we resolve ambiguities in the data by assuming that neural activity from nearby sources will tend to be similar, all else being equal. One can then ask, for an arbitrary individual's brain: given what we know about the correlational structure of other people's brains, and given the recordings we made from electrodes implanted in this person's brain, how would those recordings most likely have looked at other locations throughout this person's brain?

Generalized neural decoders for transfer learning across participants and recording modalities

“…We are currently exploring extensions of HTNet for a variety of applications such as cross-frequency coupling [94, 95], long-term state decoding [6], cross-task decoding [96], and data-driven regression [97, 98]. In addition, other decoding measures could be substituted for the Hilbert transform, including non-Fourier methods [99, 100], and more complex interpolation schemes could be used to generate the projection matrix by incorporating participant-specific cortical anatomy [101, 102]. Besides ECoG and EEG, HTNet may also be useful for generalizing across participants with stereotactic EEG or local field potential recordings [103, 104].…”

Distributed Subnetworks of Depression Defined by Direct Intracranial Neurophysiology

“…More importantly, the methods and findings provide a basis for applying similar approaches to broader depression populations and the potential for integrating such findings with non-invasive network features. While our whole-brain iEEG model was extensive in coverage, we did not have electrodes placed in all brain regions, including some regions implicated in depression [87][88][89][90][91] and the density of electrode sampling varied across brain regions leading to uncertainty in the accuracy of estimation in sparsely sampled areas 47 . We dealt with this constraint by discounting the effect of each individual node degree before running community detection and comparing network measures to a null model that accounted for overall node density.…”

“…A functional connectivity imputation technique was utilized to estimate whole-brain iEEG activity for each subject (SuperEEG, developed by Jeremy Manning and Lucy Owens) 47 . This method involved four main steps outlined in Supplementary Fig.…”

Xi rhythms: decoding neural oscillations to create full-brain high-resolution spectra parametric mapping

“…Some parameters from the spectrum components are extracted as interpretable indicators of neural oscillation, such as the oscillation amplitude, resonance frequency, bandwidth, skewness, kurtosis, slope, and relative power, say the prominence to the total power. This will allows the application of Xi Rhythms into 1) building oscillation norms over sensor and source space with large normative database; 2) prediction and differential diagnosing as biomarkers for cognitive process and mental order; 3) spatial inference with iEEG/ECOG data from recorded region to unobserved areas over the intrasubject level and the inter individual subjects level as well (Owen and Manning, 2017); 4) the full brain high resolution statistical spectrum parametric mapping may provide a norm prior for the brain magnetic-electrical inverse source imaging. Figure 1.…”

“…A multinational quantitative oscillation norm is built with the collected EEG data from Cuba, United states, Switzerland and Mexico. A full brain high resolution spectrum parametric mapping (SPM) is created thanks to the shared iEEG dataset (Frauscher et al, 2018;Owen and Manning, 2017).…”