Karim Jerbi

Good practice for conducting and reporting MEG research

Magnetoencephalographic (MEG) recordings are a rich source of information about the neural dynamics underlying cognitive processes in the brain, with excellent temporal and good spatial resolution. In recent years there have been considerable advances in MEG hardware developments and methods. Sophisticated analysis techniques are now routinely applied and continuously improved, leading to fascinating insights into the intricate dynamics of neural processes. However, the rapidly increasing level of complexity of the different steps in a MEG study make it difficult for novices, and sometimes even for experts, to stay aware of possible limitations and caveats. Furthermore, the complexity of MEG data acquisition and data analysis requires special attention when describing MEG studies in publications, in order to facilitate interpretation and reproduction of the results. This manuscript aims at making recommendations for a number of important data acquisition and data analysis steps and suggests details that should be specified in manuscripts reporting MEG studies. These recommendations will hopefully serve as guidelines that help to strengthen the position of the MEG research community within the field of neuroscience, and may foster discussion in order to further enhance the quality and impact of MEG research.

Coherent neural representation of hand speed in humans revealed by MEG imaging

The spiking activity of single neurons in the primate motor cortex is correlated with various limb movement parameters, including velocity. Recent findings obtained using local field potentials suggest that hand speed may also be encoded in the summed activity of neuronal populations. At this macroscopic level, the motor cortex has also been shown to display synchronized rhythmic activity modulated by motor behavior. Yet whether and how neural oscillations might be related to limb speed control is still poorly understood. Here, we applied magnetoencephalography (MEG) source imaging to the ongoing brain activity in subjects performing a continuous visuomotor (VM) task. We used coherence and phase synchronization to investigate the coupling between the estimated activity throughout the brain and the simultaneously recorded instantaneous hand speed. We found significant phase locking between slow (2- to 5-Hz) oscillatory activity in the contralateral primary motor cortex and time-varying hand speed. In addition, we report long-range task-related coupling between primary motor cortex and multiple brain regions in the same frequency band. The detected large-scale VM network spans several cortical and subcortical areas, including structures of the frontoparietal circuit and the cerebello–thalamo–cortical pathway. These findings suggest a role for slow coherent oscillations in mediating neural representations of hand kinematics in humans and provide further support for the putative role of long-range neural synchronization in large-scale VM integration. Our findings are discussed in the context of corticomotor communication, distributed motor encoding, and possible implications for brain–machine interfaces.

Task-related gamma-band dynamics from an intracerebral perspective: review and implications for surface EEG and MEG.

Although non‐invasive techniques provide functional activation maps at ever‐growing spatio‐temporal precision, invasive recordings offer a unique opportunity for direct investigations of the fine‐scale properties of neural mechanisms in focal neuronal populations. In this review we provide an overview of the field of intracranial Electroencephalography (iEEG) and discuss its strengths and limitations and its relationship to non‐invasive brain mapping techniques. We discuss the characteristics of invasive data acquired from implanted epilepsy patients using stereotactic‐electroencephalography (SEEG) and electrocorticography (ECoG) and the use of spectral analysis to reveal task‐related modulations in multiple frequency components. Increasing evidence suggests that gamma‐band activity (>40 Hz) might be a particularly efficient index for functional mapping. Moreover, the detection of high gamma activity may play a crucial role in bridging the gap between electrophysiology and functional imaging studies as well as in linking animal and human data. The present review also describes recent advances in real‐time invasive detection of oscillatory modulations (including gamma activity) in humans. Furthermore, the implications of intracerebral findings on future non‐invasive studies are discussed. Hum Brain Mapp, 2009.

Transient Suppression of Broadband Gamma Power in the Default-Mode Network Is Correlated with Task Complexity and Subject Performance

Task performance is associated with increased brain metabolism but also with prominent deactivation in specific brain structures known as the default-mode network (DMN). The role of DMN deactivation remains enigmatic in part because its electrophysiological correlates, temporal dynamics, and link to behavior are poorly understood. Using extensive depth electrode recordings in humans, we provide first electrophysiological evidence for a direct correlation between the dynamics of power decreases in the DMN and individual subject behavior. We found that all DMN areas displayed transient suppressions of broadband gamma (60–140 Hz) power during performance of a visual search task and, critically, we show for the first time that the millisecond range duration and extent of the transient gamma suppressions are correlated with task complexity and subject performance. In addition, trial-by-trial correlations revealed that spatially distributed gamma power increases and decreases formed distinct anticorrelated large-scale networks. Beyond unraveling the electrophysiological basis of DMN dynamics, our results suggest that, rather than indicating a mere switch to a global exteroceptive mode, DMN deactivation encodes the extent and efficiency of our engagement with the external world. Furthermore, our findings reveal a pivotal role for broadband gamma modulations in the interplay between task-positive and task-negative networks mediating efficient goal-directed behavior and facilitate our understanding of the relationship between electrophysiology and neuroimaging studies of intrinsic brain networks.

Exceeding chance level by chance: The caveat of theoretical chance levels in brain signal classification and statistical assessment of decoding accuracy

Machine learning techniques are increasingly used in neuroscience to classify brain signals. Decoding performance is reflected by how much the classification results depart from the rate achieved by purely random classification. In a 2-class or 4-class classification problem, the chance levels are thus 50% or 25% respectively. However, such thresholds hold for an infinite number of data samples but not for small data sets. While this limitation is widely recognized in the machine learning field, it is unfortunately sometimes still overlooked or ignored in the emerging field of brain signal classification. Incidentally, this field is often faced with the difficulty of low sample size. In this study we demonstrate how applying signal classification to Gaussian random signals can yield decoding accuracies of up to 70% or higher in two-class decoding with small sample sets. Most importantly, we provide a thorough quantification of the severity and the parameters affecting this limitation using simulations in which we manipulate sample size, class number, cross-validation parameters (k-fold, leave-one-out and repetition number) and classifier type (Linear-Discriminant Analysis, Naïve Bayesian and Support Vector Machine). In addition to raising a red flag of caution, we illustrate the use of analytical and empirical solutions (binomial formula and permutation tests) that tackle the problem by providing statistical significance levels (p-values) for the decoding accuracy, taking sample size into account. Finally, we illustrate the relevance of our simulations and statistical tests on real brain data by assessing noise-level classifications in Magnetoencephalography (MEG) and intracranial EEG (iEEG) baseline recordings.

Recording and analysis techniques for high-frequency oscillations

In recent years, new recording technologies have advanced such that, at high temporal and spatial resolutions, high-frequency oscillations (HFO) can be recorded in human partial epilepsy. However, because of the deluge of multichannel data generated by these experiments, achieving the full potential of parallel neuronal recordings depends on the development of new data mining techniques to extract meaningful information relating to time, frequency and space. Here, we aim to bridge this gap by focusing on up-to-date recording techniques for measurement of HFO and new analysis tools for their quantitative assessment. In particular, we emphasize how these methods can be applied, what property might be inferred from neuronal signals, and potentially productive future directions.

Exploring the electrophysiological correlates of the default-mode network with intracerebral EEG

While functional imaging studies allow for a precise spatial characterization of resting state networks, their neural correlates and thereby their fine-scale temporal dynamics remain elusive. A full understanding of the mechanisms at play requires input from electrophysiological studies. Here, we discuss human and non-human primate electrophysiological data that explore the neural correlates of the default-mode network. Beyond the promising findings obtained with non-invasive approaches, emerging evidence suggests that invasive recordings in humans will be crucial in order to elucidate the neural correlates of the brains default-mode function. In particular, we contend that stereotactic-electroencephalography, which consists of implanting multiple depth electrodes for pre-surgical evaluation in drug-resistant epilepsy, is particularly suited for this endeavor. We support this view by providing rare data from depth recordings in human posterior cingulate cortex and medial prefrontal cortex that show transient neural deactivation during task-engagement.

ELAN: a software package for analysis and visualization of MEG, EEG, and LFP signals

The recent surge in computational power has led to extensive methodological developments and advanced signal processing techniques that play a pivotal role in neuroscience. In particular, the field of brain signal analysis has witnessed a strong trend towards multidimensional analysis of large data sets, for example, single-trial time-frequency analysis of high spatiotemporal resolution recordings. Here, we describe the freely available ELAN software package which provides a wide range of signal analysis tools for electrophysiological data including scalp electroencephalography (EEG), magnetoencephalography (MEG), intracranial EEG, and local field potentials (LFPs). The ELAN toolbox is based on 25 years of methodological developments at the Brain Dynamics and Cognition Laboratory in Lyon and was used in many papers including the very first studies of time-frequency analysis of EEG data exploring evoked and induced oscillatory activities in humans. This paper provides an overview of the concepts and functionalities of ELAN, highlights its specificities, and describes its complementarity and interoperability with other toolboxes.

Simultaneous MEG and intracranial EEG recordings during attentive reading

The relationship between neural oscillations recorded at various spatial scales remains poorly understood partly due to an overall dearth of studies utilizing simultaneous measurements. In an effort to study quantitative markers of attention during reading, we performed simultaneous magnetoencephalography (MEG) and intracranial electroencephalography (iEEG) recordings in four epileptic patients. Patients were asked to attend to a specific color when presented with an intermixed series of red words and green words, with words of a given color forming a cohesive story. We analyzed alpha, beta, and gamma band oscillatory responses to the word presentation and compared the strength and spatial organization of those responses in both electrophysiological recordings. Time-frequency analysis of iEEG revealed a network of clear attention-modulated high gamma band (50-150 Hz) power increases and alpha/beta (9-25 Hz) suppressions in response to the words. In addition to analyses at the sensor level, MEG time-frequency analysis was performed at the source level using a sliding window beamformer technique. Strong alpha/beta suppressions were observed in MEG reconstructions, in tandem with iEEG effects. While the MEG counterpart of high gamma band enhancement was difficult to interpret at the sensor level in two patients, MEG time-frequency source reconstruction revealed additional activation patterns in accordance with iEEG results. Importantly, iEEG allowed us to confirm that several sources of gamma band modulation observed with MEG were indeed of cortical origin rather than EMG muscular or ocular artifact.

On MEG forward modelling using multipolar expansions

Magnetoencephalography (MEG) is a non-invasive functional imaging modality based on the measurement of the external magnetic field produced by neural current sources within the brain. The reconstruction of the underlying sources is a severely ill-posed inverse problem typically tackled using either low-dimensional parametric source models, such as an equivalent current dipole (ECD), or high-dimensional minimum-norm imaging techniques. The inability of the ECD to properly represent non-focal sources and the over-smoothed solutions obtained by minimum-norm methods underline the need for an alternative approach. Multipole expansion methods have the advantages of the parametric approach while at the same time adequately describing sources with significant spatial extent and arbitrary activation patterns. In this paper we first present a comparative review of spherical harmonic and Cartesian multipole expansion methods that can be used in MEG. The equations are given for the general case of arbitrary conductors and realistic sensor configurations and also for the special cases of spherically symmetric conductors and radially oriented sensors. We then report the results of computer simulations used to investigate the ability of a first-order multipole model (dipole and quadrupole) to represent spatially extended sources, which are simulated by 2D and 3D clusters of elemental dipoles. The overall field of a cluster is analysed using singular value decomposition and compared to the unit fields of a multipole, centred in the middle of the cluster, using subspace correlation metrics. Our results demonstrate the superior utility of the multipolar source model over ECD models in providing source representations of extended regions of activity.