Virtual brain twins: from basic neuroscience to clinical use.

Virtual brain twins are personalized, generative and adaptive brain models based on data from an individual's brain for scientific and clinical use. After a description of the key elements of virtual brain twins, we present the standard model for personalized whole-brain network models. The personalization is accomplished using a subject's brain imaging data by three means: (1) assemble cortical and subcortical areas in the subject-specific brain space; (2) directly map connectivity into the brain models, which can be generalized to other parameters; and (3) estimate relevant parameters through model inversion, typically using probabilistic machine learning. We present the use of personalized whole-brain network models in healthy ageing and five clinical diseases: epilepsy, Alzheimer's disease, multiple sclerosis, Parkinson's disease and psychiatric disorders. Specifically, we introduce spatial masks for relevant parameters and demonstrate their use based on the physiological and pathophysiological hypotheses. Finally, we pinpoint the key challenges and future directions.

Delineating epileptogenic networks using brain imaging data and personalized modeling in drug-resistant epilepsy.

Precise estimates of epileptogenic zone networks (EZNs) are crucial for planning intervention strategies to treat drug-resistant focal epilepsy. Here, we present the virtual epileptic patient (VEP), a workflow that uses personalized brain models and machine learning methods to estimate EZNs and to aid surgical strategies. The structural scaffold of the patient-specific whole-brain network model is constructed from anatomical T1 and diffusion-weighted magnetic resonance imaging. Each network node is equipped with a mathematical dynamical model to simulate seizure activity. Bayesian inference methods sample and optimize key parameters of the personalized model using functional stereoelectroencephalography recordings of patients' seizures. These key parameters together with their personalized model determine a given patient's EZN. Personalized models were further used to predict the outcome of surgical intervention using virtual surgeries. We evaluated the VEP workflow retrospectively using 53 patients with drug-resistant focal epilepsy. VEPs reproduced the clinically defined EZNs with a precision of 0.6, where the physical distance between epileptogenic regions identified by VEP and the clinically defined EZNs was small. Compared with the resected brain regions of 25 patients who underwent surgery, VEP showed lower false discovery rates in seizure-free patients (mean, 0.028) than in non-seizure-free patients (mean, 0.407). VEP is now being evaluated in an ongoing clinical trial (EPINOV) with an expected 356 prospective patients with epilepsy.

Human periventricular nodular heterotopia shows several interictal epileptic patterns and hyperexcitability of neuronal firing.

Periventricular nodular heterotopia (PNH) is a malformation of cortical development that frequently causes drug-resistant epilepsy. The epileptogenicity of ectopic neurons in PNH as well as their role in generating interictal and ictal activity is still a matter of debate. We report the first in vivo microelectrode recording of heterotopic neurons in humans. Highly consistent interictal patterns (IPs) were identified within the nodules: (1) Periodic Discharges PLUS Fast activity (PD+F), (2) Sporadic discharges PLUS Fast activity (SD+F), and (3) epileptic spikes (ES). Neuronal firing rates were significantly modulated during all IPs, suggesting that multiple IPs were generated by the same local neuronal populations. Furthermore, firing rates closely followed IP morphologies. Among the different IPs, the SD+F pattern was found only in the three nodules that were actively involved in seizure generation but was never observed in the nodule that did not take part in ictal discharges. On the contrary, PD+F and ES were identified in all nodules. Units that were modulated during the IPs were also found to participate in seizures, increasing their firing rate at seizure onset and maintaining an elevated rate during the seizures. Together, nodules in PNH are highly epileptogenic and show several IPs that provide promising pathognomonic signatures of PNH. Furthermore, our results show that PNH nodules may well initiate seizures.

A brain atlas of axonal and synaptic delays based on modelling of cortico-cortical evoked potentials.

Epilepsy presurgical investigation may include focal intracortical single-pulse electrical stimulations with depth electrodes, which induce cortico-cortical evoked potentials at distant sites because of white matter connectivity. Cortico-cortical evoked potentials provide a unique window on functional brain networks because they contain sufficient information to infer dynamical properties of large-scale brain connectivity, such as preferred directionality and propagation latencies. Here, we developed a biologically informed modelling approach to estimate the neural physiological parameters of brain functional networks from the cortico-cortical evoked potentials recorded in a large multicentric database. Specifically, we considered each cortico-cortical evoked potential as the output of a transient stimulus entering the stimulated region, which directly propagated to the recording region. Both regions were modelled as coupled neural mass models, the parameters of which were estimated from the first cortico-cortical evoked potential component, occurring before 80 ms, using dynamic causal modelling and Bayesian model inversion. This methodology was applied to the data of 780 patients with epilepsy from the F-TRACT database, providing a total of 34 354 bipolar stimulations and 774 445 cortico-cortical evoked potentials. The cortical mapping of the local excitatory and inhibitory synaptic time constants and of the axonal conduction delays between cortical regions was obtained at the population level using anatomy-based averaging procedures, based on the Lausanne2008 and the HCP-MMP1 parcellation schemes, containing 130 and 360 parcels, respectively. To rule out brain maturation effects, a separate analysis was performed for older (>15 years) and younger patients (<15 years). In the group of older subjects, we found that the cortico-cortical axonal conduction delays between parcels were globally short (median = 10.2 ms) and only 16% were larger than 20 ms. This was associated to a median velocity of 3.9 m/s. Although a general lengthening of these delays with the distance between the stimulating and recording contacts was observed across the cortex, some regions were less affected by this rule, such as the insula for which almost all efferent and afferent connections were faster than 10 ms. Synaptic time constants were found to be shorter in the sensorimotor, medial occipital and latero-temporal regions, than in other cortical areas. Finally, we found that axonal conduction delays were significantly larger in the group of subjects younger than 15 years, which corroborates that brain maturation increases the speed of brain dynamics. To our knowledge, this study is the first to provide a local estimation of axonal conduction delays and synaptic time constants across the whole human cortex in vivo, based on intracerebral electrophysiological recordings.

Long-term deep intracerebral microelectrode recordings in patients with drug-resistant epilepsy: Proposed guidelines based on 10-year experience.

PURPOSE: Human neuronal activity, recorded in vivo from microelectrodes, may offer valuable insights into physiological mechanisms underlying human cognition and pathophysiological mechanisms of brain diseases, in particular epilepsy. Continuous and long-term recordings are necessary to monitor non predictable pathological and physiological activities like seizures or sleep. Because of their high impedance, microelectrodes are more sensitive to noise than macroelectrodes. Low noise levels are crucial to detect action potentials from background noise, and to further isolate single neuron activities. Therefore, long-term recordings of multi-unit activity remains a challenge. We shared here our experience with microelectrode recordings and our efforts to reduce noise levels in order to improve signal quality. We also provided detailed technical guidelines for the connection, recording, imaging and signal analysis of microelectrode recordings. RESULTS: During the last 10 years, we implanted 122 bundles of Behnke-Fried hybrid macro-microelectrodes, in 56 patients with pharmacoresistant focal epilepsy. Microbundles were implanted in the temporal lobe (74%), as well as frontal (15%), parietal (6%) and occipital (5%) lobes. Low noise levels depended on our technical setup. The noise reduction was mainly obtained after electrical insulation of the patient's recording room and the use of a reinforced microelectrode model, reaching median root mean square values of 5.8 µV. Seventy percent of the bundles could record multi-units activities (MUA), on around 3 out of 8 wires per bundle and for an average of 12 days. Seizures were recorded by microelectrodes in 91% of patients, when recorded continuously, and MUA were recorded during seizures for 75 % of the patients after the insulation of the room. Technical guidelines are proposed for (i) electrode tails manipulation and protection during surgical bandage and connection to both clinical and research amplifiers, (ii) electrical insulation of the patient's recording room and shielding, (iii) data acquisition and storage, and (iv) single-units activities analysis. CONCLUSIONS: We progressively improved our recording setup and are now able to record (i) microelectrode signals with low noise level up to 3 weeks duration, and (ii) MUA from an increased number of wires . We built a step by step procedure from electrode trajectory planning to recordings. All these delicate steps are essential for continuous long-term recording of units in order to advance in our understanding of both the pathophysiology of ictogenesis and the neuronal coding of cognitive and physiological functions.

Human resting-state EEG and radiofrequency GSM mobile phone exposure: the impact of the individual alpha frequency.

PURPOSE: With the extensive use of mobile phone (MP), several studies have been realized to investigate the effects of radiofrequency electromagnetic fields (RF-EMF) exposure on brain activity at rest via electroencephalography (EEG), and the most consistent effect has been seen on the alpha band power spectral density (PSD). However, some studies reported an increase or a decrease of the PSD, while others showed no effect. It has been suggested that these differences might partly be due to a variability of the physiological state of the brain between subjects. So, the aim of this study was to investigate the alpha band modulation, exploring the impact of the alpha band frequency ranges applied in the PSD analysis. MATERIALS AND METHODS: Twenty-one healthy volunteers took part to the study with a double-blind, randomized and counterbalanced crossover design, during which eyes-open (EO) and eyes-closed (EC) resting-state EEG was recorded. The exposure system was a sham or a real GSM (global system for mobile) 900 MHz MP (pulse modulated at 217 Hz, mean power of 250 mW and 2 W peak, with a maximum specific absorption rate of 0.70 W/kg on 1 g tissue). The experimental protocol presented a baseline recording phase without MP exposure, an exposure phase during which the exposure system was placed against the left ear, and the post-exposure phase without MP. EEG data from baseline and exposure phases were analyzed and PSD was computed for the alpha band in the fixed range of 8-12 Hz and for the individual alpha band frequency range (IAF). RESULTS: Results showed a trend in decrease or increase of EEG power of both alpha oscillations during exposure in relation to EC and EO recording conditions, respectively, but not reaching statistical significance. Findings did not provide evidence for a different sensitivity to RF-EMF MP related to individual variability in the frequency of the alpha band. CONCLUSION: In conclusion, these results did not show alpha band activity modulation during resting-state under RF-EMF. It might be argued the need of a delay after the exposure in order to appreciate an EEG spectral power modulation related to RF-EMF exposure.

Modulation of magnetoencephalography alpha band activity by radiofrequency electromagnetic field depicted in sensor and source space.

Several studies reported changes in spontaneous electroencephalogram alpha band activity related to radiofrequency electromagnetic fields, but findings showed both an increase and a decrease of its spectral power or no effect. Here, we studied the alpha band modulation after 900 MHz mobile phone radiofrequency exposure and localized cortical regions involved in these changes, via a magnetoencephalography (MEG) protocol with healthy volunteers in a double-blind, randomized, counterbalanced crossover design. MEG was recorded during eyes open and eyes closed resting-state before and after radiofrequency exposure. Potential confounding factors, known to affect alpha band activity, were assessed as control parameters to limit bias. Entire alpha band, lower and upper alpha sub-bands MEG power spectral densities were estimated in sensor and source space. Biochemistry assays for salivary biomarkers of stress (cortisol, chromogranin-A, alpha amylase), heart rate variability analysis and high-performance liquid chromatography for salivary caffeine concentration were realized. Results in sensor and source space showed a significant modulation of MEG alpha band activity after the radiofrequency exposure, with different involved cortical regions in relation to the eyes condition, probably because of different attention level with open or closed eyes. None of the control parameters reported a statistically significant difference between experimental sessions.

Comparison of two integration methods for dynamic causal modeling of electrophysiological data.

Dynamic causal modeling (DCM) is a methodological approach to study effective connectivity among brain regions. Based on a set of observations and a biophysical model of brain interactions, DCM uses a Bayesian framework to estimate the posterior distribution of the free parameters of the model (e.g. modulation of connectivity) and infer architectural properties of the most plausible model (i.e. model selection). When modeling electrophysiological event-related responses, the estimation of the model relies on the integration of the system of delay differential equations (DDEs) that describe the dynamics of the system. In this technical note, we compared two numerical schemes for the integration of DDEs. The first, and standard, scheme approximates the DDEs (more precisely, the state of the system, with respect to conduction delays among brain regions) using ordinary differential equations (ODEs) and solves it with a fixed step size. The second scheme uses a dedicated DDEs solver with adaptive step sizes to control error, making it theoretically more accurate. To highlight the effects of the approximation used by the first integration scheme in regard to parameter estimation and Bayesian model selection, we performed simulations of local field potentials using first, a simple model comprising 2 regions and second, a more complex model comprising 6 regions. In these simulations, the second integration scheme served as the standard to which the first one was compared. Then, the performances of the two integration schemes were directly compared by fitting a public mismatch negativity EEG dataset with different models. The simulations revealed that the use of the standard DCM integration scheme was acceptable for Bayesian model selection but underestimated the connectivity parameters and did not allow an accurate estimation of conduction delays. Fitting to empirical data showed that the models systematically obtained an increased accuracy when using the second integration scheme. We conclude that inference on connectivity strength and delay based on DCM for EEG/MEG requires an accurate integration scheme.

Amygdala processing of social cues from faces: an intracrebral EEG study.

The amygdala is a key structure for monitoring the relevance of environmental stimuli. Yet, little is known about the dynamics of its response to primary social cues such as gaze and emotion. Here, we examined evoked amygdala responses to gaze and facial emotion changes in five epileptic patients with intracerebral electrodes. Patients first viewed a neutral face that would then convey social cues: it turned either happy or fearful with or without gaze aversion. This social cue was followed by a laterally presented target, the detection of which was faster if it appeared in a location congruent with the averted gaze direction. First, we observed pronounced evoked amygdala potentials to the initial neutral face. Second, analysis of the evoked responses to the cue showed an early effect of gaze starting at 123 ms in the right amygdala. Differential effects of fearful vs happy valence were individually present but more variable in time and therefore not observed at group-level. Our study is the first to demonstrate such an early effect of gaze in the amygdala, in line with its particular behavioral relevance in the spatial attention task.

Radiofrequency signal affects alpha band in resting electroencephalogram.

The aim of the present work was to investigate the effects of the radiofrequency (RF) electromagnetic fields (EMFs) on human resting EEG with a control of some parameters that are known to affect alpha band, such as electrode impedance, salivary cortisol, and caffeine. Eyes-open and eyes-closed resting EEG data were recorded in 26 healthy young subjects under two conditions: sham exposure and real exposure in double-blind, counterbalanced, crossover design. Spectral power of EEG rhythms was calculated for the alpha band (8-12 Hz). Saliva samples were collected before and after the study. Salivary cortisol and caffeine were assessed by ELISA and HPLC, respectively. The electrode impedance was recorded at the beginning of each run. Compared with the sham session, the exposure session showed a statistically significant (P < 0.0001) decrease of the alpha band spectral power during closed-eyes condition. This effect persisted in the postexposure session (P < 0.0001). No significant changes were detected in electrode impedance, salivary cortisol, and caffeine in the sham session compared with the exposure one. These results suggest that GSM-EMFs of a mobile phone affect the alpha band within spectral power of resting human EEG.

Behavioral evidence for differences in social and non-social category learning.

When meeting someone for the very first time one spontaneously categorizes the seen person on the basis of his/her appearance. Categorization is based on the association between some physical features and category labels that can be social (character trait…) or non-social (tall, thin). Surprisingly little is known about how such associations are formed, particularly in the social domain. Here, we aimed at testing whether social and non-social category learning may be dissociated. We presented subjects with a large number of faces that had to be rated according to social or non-social labels, and induced an association between a facial feature (inter-eye distance) and the category labels using two different procedures. In a first experiment, we used a feedback procedure to reinforce the association; behavioral measures revealed an association between the physical feature manipulated and abstract non-social categories, while no evidence for an association with social labels could be found. In a second experiment, we used passive exposure to the association between physical features and labels; we obtained behavioral evidence for learning of both social and non-social categories. These results support the view of the specificity of social category learning; they suggest that social categories are best acquired through unsupervised procedures that can be considered as a simplified proxy for group transmission.

Oscillatory Brain Correlates of Live Joint Attention: A Dual-EEG Study.

Joint attention consists in following another's gaze onto an environmental object, which leads to the alignment of both subjects' attention onto this object. It is a fundamental mechanism of non-verbal communication, and it is essential for dynamic, online, interindividual synchronization during interactions. Here we aimed at investigating the oscillatory brain correlates of joint attention in a face-to-face paradigm where dyads of participants dynamically oriented their attention toward the same or different objects during joint and no-joint attention periods respectively. We also manipulated task instruction: in socially driven instructions, the participants had to follow explicitly their partner's gaze, while in color-driven instructions, the objects to be looked at were designated at by their color so that no explicit gaze following was required. We focused on oscillatory activities in the 10 Hz frequency range, where parieto-occipital alpha and the centro-parietal mu rhythms have been described, as these rhythms have been associated with attention and social coordination processes respectively. We tested the hypothesis of a modulation of these oscillatory activities by joint attention. We used dual-EEG to record simultaneously the brain activities of the participant dyads during our live, face-to-face joint attention paradigm. We showed that joint attention periods - as compared to the no-joint attention periods - were associated with a decrease of signal power between 11 and 13 Hz over a large set of left centro-parieto-occipital electrodes, encompassing the scalp regions where alpha and mu rhythms have been described. This 11-13 Hz signal power decrease was observed independently of the task instruction: it was similar when joint versus no-joint attention situations were socially driven and when they were color-driven. These results are interpreted in terms of the processes of attention mirroring, social coordination, and mutual attentiveness associated with joint attention state.

Behavioural and magnetoencephalographic evidence for the interaction between semantic and episodic memory in healthy elderly subjects.

The relationship between episodic and semantic memory systems has long been debated. Some authors argue that episodic memory is contingent on semantic memory (Tulving 1984), while others postulate that both systems are independent since they can be selectively damaged (Squire 1987). The interaction between these memory systems is particularly important in the elderly, since the dissociation of episodic and semantic memory defects characterize different aging-related pathologies. Here, we investigated the interaction between semantic knowledge and episodic memory processes associated with faces in elderly subjects using an experimental paradigm where the semantic encoding of famous and unknown faces was compared to their episodic recognition. Results showed that the level of semantic awareness of items affected the recognition of those items in the episodic memory task. Event-related magnetic fields confirmed this interaction between episodic and semantic memory: ERFs related to the old/new effect during the episodic task were markedly different for famous and unknown faces. The old/new effect for famous faces involved sustained activities maximal over right temporal sensors, showing a spatio-temporal pattern partly similar to that found for famous versus unknown faces during the semantic task. By contrast, an old/new effect for unknown faces was observed on left parieto-occipital sensors. These findings suggest that the episodic memory for famous faces activated the retrieval of stored semantic information, whereas it was based on items' perceptual features for unknown faces. Overall, our results show that semantic information interfered markedly with episodic memory processes and suggested that the neural substrates of these two memory systems overlap.

Early influence of prior experience on face perception.

Inferring someone's personality from his or her photograph is a pervasive and automatic behavior that takes place even if no reliable information about one's character can be derived solely from facial features. This illustrates nicely the idea that perception is not a passive process, but rather an active combination of current sensory inputs with endogenous knowledge derived from prior experience. To understand how and when neural responses to faces can be modulated by prior experience, we recorded magneto-encephalographic (MEG) responses to new faces, before and after subjects were exposed for a short period of 15-20 min to an experimentally induced association between a facial feature (inter-eye distance) and a response (personality judgment). In spite of the absence of any observable response bias following such a short reinforcement phase, our experimental manipulation influenced neural responses to faces as early as 60-85 ms. Source localization of magneto-encephalographic signals, confirmed by intracranial recordings, suggests that prior experience modulates early neural processing along two initially independent neural routes, one initiated in an anterior system that includes the orbitofrontal cortex and the temporal poles, and the second one involving face-sensitive regions in the ventral visual pathway. The two routes are both active as early as 60 ms but engage in reciprocal interactions only later, between 135 and 160 ms. These experimental findings support recent models assuming the existence of a fast anterior pathway activated in parallel with the ventral visual system which would link prior experience with current sensory inputs.