Search

Advances in functional magnetic resonance spectroscopy (fMRS) have enabled the quantification of activity-dependent changes in neurotransmitter concentrations in vivo. However, the physiological basis of the large changes in GABA and glutamate observed by fMRS (>10%) over short time scales of less than a minute remain unclear as such changes cannot be accounted for by known synthesis or degradation metabolic pathways. Instead, it has been hypothesized that fMRS detects shifts in neurotransmitter concentrations as they cycle from presynaptic vesicles, where they are largely invisible, to extracellular and cytosolic pools, where they are detectable. The present paper uses a computational modelling approach to demonstrate the viability of this hypothesis. A new mean-field model of the neural mechanisms generating the fMRS signal in a cortical voxel is derived. The proposed macroscopic mean-field model is based on a microscopic description of the neurotransmitter dynamics at the level of the synapse. Specifically, GABA and glutamate are assumed to cycle between three metabolic pools: packaged in the vesicles; active in the synaptic cleft; and undergoing recycling and repackaging in the astrocytic or neuronal cytosol. Computational simulations from the model are used to generate predicted changes in GABA and glutamate concentrations in response to different types of stimuli including pain, vision, and electric current stimulation. The predicted changes in the extracellular and cytosolic pools corresponded to those reported in empirical fMRS data. Furthermore, the model predicts a selective control mechanism of the GABA/glutamate relationship, whereby inhibitory stimulation reduces both neurotransmitters, whereas excitatory stimulation increases glutamate and decreases GABA. The proposed model bridges between neural dynamics and fMRS and provides a mechanistic account for the activity-dependent changes in the glutamate and GABA fMRS signals. Lastly, these results indicate that echo-time may be an important timing parameter that can be leveraged to maximise fMRS experimental outcomes.

The relationship between structural and functional brain networks has been characterised as complex: the two networks mirror each other and show mutual influence but they also diverge in their organisation. This work explored whether a combination of structural and functional connectivity can improve predictive models of cognitive performance. Principal Component Analysis (PCA) was first applied to cognitive data from the Human Connectome Project to identify components reflecting five cognitive domains: Executive Function, Self-regulation, Language, Encoding and Sequence Processing. A Principal Component Regression (PCR) approach was then used to fit predictive models of each cognitive domain based on structural (SC), functional (FC) or combined structural-functional (CC) connectivity. Self-regulation, Encoding and Sequence Processing were best modelled by FC, whereas Executive Function and Language were best modelled by CC. The present study demonstrates that integrating structural and functional connectivity can help predict cognitive performance, but that the added explanatory value may be (cognitive) domain-specific. Implications of these results for studies of the brain basis of cognition in health and disease are discussed.HighlightsWe assessed the relationship between cognitive domains and structural, functional and combined structural-functional connectivity.We found that Executive Function and Language components were best predicted by combined models of functional and structural connectivity.Self-regulation, Encoding and Sequence Processing were best predicted by functional connectivity alone.Our findings provide insight into separable contributions of functional, structural and combined connectivity to different cognitive domains.

Understanding the role of neurotransmitters glutamate and GABA during normal and abnormal brain function and under external stimulation in humans are critical neuroscientific and clinical goals. The recent development of functional 1H-Magnetic resonance spectroscopy (fMRS) has allowed us to study neuro-transmitter activity in vivo for the first time. However, the physiological basis of the observed fMRS signal remains unclear. It has been proposed that fMRS detects shifts in metabolite concentrations as they move from presynaptic vesicles, where they are largely invisible, to extracellular and cytosolic pools, where they are visible.Here we bridge the gap between neural dynamics and fMRS by developing a mean-field model to link the neurotransmitter dynamics at the microscopic-level to the macroscopic-level imaging measurements. GABA and glutamate are described as cycling between three metabolic pools: in the vesicles; active in the cleft; or undergoing recycling in the astrocytic or neuronal cytosol. We interrogate the model by applying a current to manipulate the mean membrane potential and firing rate of the neural populations.We find that by disregarding the contribution from the vesicular pool, our model predicts activity-dependent changes in the MRS signal, which are consistent with reported empirical findings. Further, we show that current magnitude and direction has a selective effect on the GABA/glutamate-MRS signal: inhibitory stimulation leads to reduction of both metabolites, whereas excitatory stimulation leads to increased glutamate and decreased GABA. In doing so, we link neural dynamics and fMRS and provide a mechanistic account for the activity-dependent change in the observed MRS signal.Key Points SummaryThe recent development of functional 1H-Magnetic resonance spectroscopy (fMRS) has allowed us to study neurotransmitter activity in vivo for the first time in humans. However, the physiological basis of the observed fMRS signal is unclear.It has been proposed that fMRS detects shifts in metabolite concentrations as they move from presynaptic vesicles, where they are largely invisible to MRS, to extracellular and cytosolic pools, where they are visible to MRS.We test this hypothesis using a mean field model which links the neural dynamics of neurotransmitters at the microscopic-level to the macroscopic-level imaging measurements obtained in experimental studies.By disregarding activity in the vesicular pool, our model can generate activity-dependent changes in the MRS signal in response to stimulation which are consistent with experimental findings in the literature.We provide a mechanistic account for the activity-dependent change in observed neurotransmitter concentrations using MRS.

One-third of the population in the UK and worldwide struggle with chronic pain. Entraining brain alpha activity through noninvasive visual stimulation has been shown to reduce experimental pain in healthy volunteers. Neural oscillations entrainment offers a potential noninvasive and nonpharmacological intervention for patients with chronic pain, which can be delivered in the home setting and has the potential to reduce use of medications. However, evidence supporting its use in patients with chronic pain is lacking. This study explores whether (a) alpha entrainment increase alpha power in patients and (b) whether this increase in alpha correlates with analgesia. In total, 28 patients with chronic pain sat in a comfortable position and underwent 4-min visual stimulation using customised goggles at 10 Hz (alpha) and 7 Hz (control) frequency blocks in a randomised cross-over design. 64-channel electroencephalography and 11-point numeric rating scale pain intensity and pain unpleasantness scores were recorded before and after stimulation. Electroencephalography analysis revealed frontal alpha power was significantly higher when stimulating at 10 Hz when compared to 7 Hz. There was a significant positive correlation between increased frontal alpha and reduction in pain intensity (r = 0.33; P

Summary Frequency-dependent reorganization of the primary somatosensory cortex, together with perceptual changes, arises following repetitive sensory stimulation. Here, we investigate the role of GABA in this process. We co-stimulated two finger tips and measured GABA and Glx using magnetic resonance (MR) spectroscopy at the beginning and end of the stimulation. Participants performed a perceptual learning task before and after stimulation. There were 2 sessions with stimulation frequency either at or above the resonance frequency of the primary somatosensory cortex (23 and 39 Hz, respectively). Perceptual learning occurred following above resonance stimulation only, while GABA reduced during this condition. Lower levels of early GABA were associated with greater perceptual learning. One possible mechanism underlying this finding is that cortical disinhibition “unmasks” lateral connections within the cortex to permit adaptation to the sensory environment. These results provide evidence in humans for a frequency-dependent inhibitory mechanism underlying learning and suggest a mechanism-based approach for optimizing neurostimulation frequency., Graphical Abstract, Highlights • In the context of repetitive sensory stimulation, GABA release is frequency dependent • Stimulating above the resonance frequency of the somatosensory cortex reduces GABA • Perceptual learning is associated with a reduction in GABA • Early GABA reduction opens a window for plasticity and learning, Molecular Neuroscience; Systems Neuroscience; Cognitive Neuroscience

There has been an increasing interest in examining organisational principles of the cerebral cortex (and subcortical regions) using different MRI features such as structural or functional connectivity. Despite the widespread interest, introductory tutorials on the underlying technique targeted for the novice neuroimager are sparse in the literature. Articles that investigate various “neural gradients” (for example based on region studied “cortical gradients,” “cerebellar gradients,” “hippocampal gradients” etc … or feature of interest “functional gradients,” “cytoarchitectural gradients,” “myeloarchitectural gradients” etc …) have increased in popularity. Thus, we believe that it is opportune to discuss what is generally meant by “gradient analysis”. We introduce basics concepts in graph theory, such as graphs themselves, the degree matrix, and the adjacency matrix. We discuss how one can think about gradients of feature similarity (the similarity between timeseries in fMRI, or streamline in tractography) using graph theory and we extend this to explore such gradients across the whole MRI scale; from the voxel level to the whole brain level. We proceed to introduce a measure for quantifying the level of similarity in regions of interest. We propose the term “the Vogt-Bailey index” for such quantification to pay homage to our history as a brain mapping community. We run through the techniques on sample datasets including a brain MRI as an example of the application of the techniques on real data and we provide several appendices that expand upon details. To maximise intuition, the appendices contain a didactic example describing how one could use these techniques to solve a particularly pernicious problem that one may encounter at a wedding. Accompanying the article is a tool, available in both MATLAB and Python, that enables readers to perform the analysis described in this article on their own data. We refer readers to the graphical abstract as an overview of the analysis pipeline presented in this work., Graphical abstract The two basic algorithms to compute the VB Index and the principal gradient. The algorithm on the left depicts a searchlight algorithm that identifies local borders. The algorithm on the right (yellow route) depicts the creation of a gradient map and single VB Index for the whole brain or (inclusion of orange route) multiple clusters.Image 1, Highlights • This article provides a beginner friendly review of the main steps for creating gradient maps. • We introduce a new index for quantifying mesoscopic gradients (the “VB Index”). • We use historical background to highlight the need for such quantification. • Software for computing gradient maps and the VB Index is made available.

The hub-and-spoke model of semantic cognition seeks to reconcile embodied views of a fully distributed semantic network with patient evidence, primarily from semantic dementia, who demonstrate modality-independent conceptual deficits associated with atrophy centred on the ventrolateral anterior temporal lobe. The proponents of this model have recently suggested that the temporal cortex is a graded representational space where concepts become less linked to a specific modality as they are processed farther away from primary and secondary sensory cortices and towards the ventral anterior temporal lobe. To explore whether there is evidence that the connectivity patterns of the temporal lobe converge in its ventral anterior end the current study uses three dimensional Laplacian eigenmapping, a technique that allows visualisation of similarity in a low dimensional space. In this space similarity is encoded in terms of distances between data points. We found that the ventral and anterior temporal lobe is in a unique position of being at the centre of mass of the data pointswithin the connective similarity space. This can be interpreted as the area where the connectivity profiles of all other temporal cortex voxels converge. This study is the first to explicitly investigate the pattern of connectivity and thus provides the missing link in the evidence that the ventral anterior temporal lobe can be considered a multi-modal graded hub., peer-reviewed

We consider the detection and characterisation of brain state transitions, based on ongoing Magneto and Electroencephalography (M/EEG). Here a brain state represents a specific brain dynamical regime or mode of operation, which produces a characteristic quasi-stable pattern of activity at topography, sources or network levels. These states and their transitions over time can reflect fundamental computational properties of the brain, shaping human behaviour and brain function. The Hidden Markov Model (HMM) has emerged as a useful model-based approach for uncovering the hidden dynamics of brain state transitions based on observed data. However, the Geometric distribution of state duration (dwell time) implicit in HMM places highest probability on very short durations, which makes it inappropriate for the accurate modelling of brain states in M/EEG. We propose using Hidden Semi Markov Models (HSMM), a generalisation of HMM that models the brain state duration distribution explicitly. We present a Bayesian formulation of HSMM and use the Variational Bayes framework to efficiently estimate the HSMM parameters, the number of brain states and select among alternative brain state duration distributions. We assess HSMM performance against HMM on simulated data and demonstrate that the accurate modelling of state duration is paramount for accurately and robustly modelling non-Markovian EEG brain state features. Finally, we used actual resting-state EEG data to demonstrate the approach in practice and conclude that it provides a flexible parameterised framework that permits closer interrogation of possible generative mechanisms.

We analyze the functional significance of different event-related potentials (ERPs) as electrophysiological indices of face perception and face recognition, according to cognitive and neurofunctional models of face processing. Initially, the processing of faces seems to be supported by early extrastriate occipital cortices and revealed by modulations of the occipital P1. This early response is thought to reflect the detection of certain primary structural aspects indicating the presence grosso modo of a face within the visual field.The posterior-temporal N170 is more sensitive to the detection of faces as complex-structured stimuli and, therefore, to the presence of its distinctive organizational characteristics prior to within-category identification. In turn, the relatively late and probably more rostrally generated N250r and N400-like responses might respectively indicate processes of access and retrieval of face-related information, which is stored in long-term memory (LTM). New methods of analysis of electrophysiological and neuroanatomical data, namely, dynamic causal modeling, single-trial and time-frequency analyses, are highly recommended to advance in the knowledge of those brain mechanisms concerning face processing, This work was supported by “Ministerio de Economía y Competitividad” (Spain, I + D + I National Programme, PSI2013-46007-P)

Frequency-dependent plasticity (FDP) describes adaptation at the synapse in response to stimulation at different frequencies. Its consequence on the structure and function of cortical networks is unknown. We tested whether cortical “resonance,” favorable stimulation frequencies at which the sensory cortices respond maximally, influenced the impact of FDP on perception, functional topography, and connectivity of the primary somatosensory cortex using psychophysics and functional imaging (fMRI). We costimulated two digits on the hand synchronously at, above, or below the resonance frequency of the somatosensory cortex, and tested subjects’ accuracy and speed on tactile localization before and after costimulation. More errors and slower response times followed costimulation at above- or below-resonance, respectively. Response times were faster after at-resonance costimulation. In the fMRI, the cortical representations of the two digits costimulated above-resonance shifted closer, potentially accounting for the poorer performance. Costimulation at-resonance did not shift the digit regions, but increased the functional coupling between them, potentially accounting for the improved response time. To relate these results to synaptic plasticity, we simulated a network of oscillators incorporating Hebbian learning. Two neighboring patches embedded in a cortical sheet, mimicking the two digit regions, were costimulated at different frequencies. Network activation outside the stimulated patches was greatest at above-resonance frequencies, reproducing the spread of digit representations seen with fMRI. Connection strengths within the patches increased following at-resonance costimulation, reproducing the increased fMRI connectivity. We show that FDP extends to the cortical level and is influenced by cortical resonance.

Recent findings indicate that phase-amplitude coupling between neuronal oscillations in the theta- (3-6 Hz) and the gamma-band (30-100 Hz) plays a functional role in memory processes. Here, using electroencephalography, we provide further evidence for coupling between prefrontal theta and parietal gamma during successful memory retrieval in the human brain. In a pictorial recognition task, the coupling between prefrontal theta phase and parietal gamma amplitude was quantified using the modulation index, 100-1500 ms after stimulus onset. Results show an increased coupling for remembered, as opposed to forgotten and new stimuli (i.e. a "recognition effect" and an "old/new effect"). Phase-amplitude coupling between the prefrontal theta phase and posterior gamma amplitudes is hypothesized to reflect long range communication between prefrontal control processes and the activation of posterior object representations accompanying mnemonic processing.

The N400 component is commonly associated with the detection of linguistic incongruity. A few studies have shown that the N400 can also be elicited by non-linguistic stimuli. Different spatiotemporal patterns were observed between the typical Linguistic N400 and the Non-linguistic N400, suggesting distinct brain generators. The aim of this study was to investigate the presence of an N400 in response to linguistic and non-linguistic stimuli, and to specify anatomical sources of both N400s using a novel analysis method: the Bayesian Model Averaging (BMA) distributed source model. Picture-word and environmental soundpicture associations, either congruent or incongruent, were presented to ten young healthy adults while highdensity ERP recordings were made. Standard electrophysiological analyses confirmed that the N400 was not specific to linguistic incongruity but was also elicited by environmental sound-picture incongruities. Different topographic distributions were obtained for the Linguistic N400 and Non-linguistic N400. BMA analysis showed that the Linguistic N400 generators were mostly located in the left superior temporal gyrus, whereas the sources of the Non-linguistic N400 were identified mostly in the right middle and superior temporal gyri. Detection of linguistic incongruities recruited cerebral areas commonly associated with language processing, whereas non-linguistic incongruities recruited right cerebral regions usually associated with auditory processing of non-linguistic stimuli. The Linguistic and Non-linguistic N400s appear to be elicited by similar cognitive mechanisms assumed by different cerebral areas depending on the type of material to be processed. The present findings support the existence of parallel pathways for the processing of linguistic and non-linguistic incongruities.

The role of induced gamma-band responses (iGBRs) in the human electroencephalogram (EEG) is a controversial topic. On the one hand, iGBRs have been associated with neuronal activity reflecting the (re-)activation of cortical object representations. On the other hand, it was shown that miniature saccades (MSs) lead to high-frequency artifacts in the EEG that can mimic cortical iGBRs. We recorded EEG and eye movements simultaneously while participants were engaged in a combined repetition priming and object recognition experiment. MS rates were mainly modulated by object familiarity in a time window from 100 to 300 ms after stimulus onset. In contrast, artifact-corrected iGBRs were sensitive to object repetition and object familiarity in a prolonged time window. EEG source analyses revealed that stimulus repetitions modulated iGBRs in temporal and occipital cortex regions while familiarity was associated with activity in parieto-occipital regions. These results are in line with neuroimaging studies employing functional magnetic resonance imaging or magnetoencephalography. We conclude that MSs reflect early mechanisms of visual perception while iGBRs mirror the activation of cortical networks representing a perceived object.

The remarkable capabilities displayed by humans in making sense of an overwhelming amount of sensory information cannot be explained easily if perception is viewed as a passive process. Current theoretical and computational models assume that to achieve meaningful and coherent perception, the human brain must anticipate upcoming stimulation. But how are upcoming stimuli predicted in the brain? We unmasked the neural representation of a prediction by omitting the predicted sensory input. Electrophysiological brain signals showed that when a clear prediction can be formulated, the brain activates a template of its response to the predicted stimulus before it arrives to our senses.

Although previous studies have established that successful memory encoding is associated with increased synchronization of theta-band and gamma-band oscillations, it is unclear if there is a functional relationship between oscillations in these frequency bands. Using scalp-recorded EEG in healthy human participants, we demonstrate that cross-frequency coupling between frontal theta phase and posterior gamma power is enhanced during the encoding of visual stimuli which participants later on remember versus items which participants subsequently forget ("subsequent memory effect," SME). Conventional wavelet analyses and source localizations revealed SMEs in spectral power of theta-, alpha-, and gamma-band. Successful compared to unsuccessful encoding was reflected in increased theta-band activity in right frontal cortex as well as increased gamma-band activity in parietal-occipital regions. Moreover, decreased alpha-band activity in prefrontal and occipital cortex was also related to successful encoding. Overall, these findings support the idea that during the formation of new memories frontal cortex regions interact with cortical representations in posterior areas.

Background There is amassing evidence that risky decision-making in bipolar disorder is related to reward-based differences in frontostriatal regions. However, the roles of early attentional and later cognitive processes remain unclear, limiting theoretical understanding and development of targeted interventions. Methods Twenty euthymic bipolar disorder and 19 matched control participants played a Roulette task in which they won and lost money. Event-related potentials and source analysis were used to quantify predominantly sensory-attentional (N1), motivational salience (feedback-related negativities [FRN]), and cognitive appraisal (P300) stages of processing. We predicted that the bipolar disorder group would show increased N1, consistent with increased attentional orienting, and reduced FRN, consistent with a bias to perceive outcomes more favorably. Results As predicted, the bipolar disorder group showed increased N1 and reduced FRN but no differences in P300. N1 amplitude was additionally associated with real-life risk taking, and N1 source activity was reduced in visual cortex but increased activity in precuneus, frontopolar, and premotor cortex, compared to those of controls. Conclusions These findings demonstrate an early attentional bias to reward that potentially drives risk taking by priming approach behavior and elevating reward salience in the frontostriatal pathway. Although later cognitive appraisals of these inputs may be relatively intact in remission, interventions targeting attention orienting may also be effective in long-term reduction of relapse.

The automatic processing of deviances from the temporal context of the visual environment has become an important topic in visual cognitive sciences, which is often investigated using the visual mismatch negativity (vMMN). This event-related potential (ERP) component is elicited by an irregular stimulus (e.g., a red disc) presented in a series of stimuli (e.g., green discs) comprising a temporal regularity (e.g., colour repetition). We determined the influence of lower- vs. upper-hemifield presentation of the irregular stimulus on the vMMN while using whole-field stimulus displays controlling for sustained shifts in spatial attention. Deviances presented in the lower hemifield elicited a larger vMMN than the ones presented in the upper hemifield at a latency of 200-280ms. However, this asymmetry was preceded by deviance-related hemifield effects already emerging at an earlier latency (110-150ms), where upper-hemifield deviances elicited a positive potential but lower-hemifield deviances did not. With variable resolution electromagnetic tomography (VARETA) early deviance-related activity was localised to retinotopically organised regions of the visual cortex (BA 17/18) and vMMN-sources were localised to the middle/superior occipital gyrus, to higher areas along the temporal visual stream, but also to BA 17/18. We argue that the upper/lower-hemifield vMMN asymmetry relies at least partially on the hemifield-dependent differential sensitivity of early deviance-related activity generated in retinotopically organised regions of the visual cortex. However, a superior automatic processing of deviances presented in the lower visual hemifield may also contribute to the effect.

Conventional neuroscientific methods are inadequate for separating the brain responses related to the simultaneous processing of different parts of a natural scene. In the present human electroencephalogram (EEG) study, we overcame this limitation by tagging concurrently presented backgrounds and objects with different presentation frequencies. As a result, background and object elicited different steady-state visual evoked potentials (SSVEPs), which were separately quantified in the frequency domain. We analyzed the effects of semantic consistency and inconsistency between background and object on SSVEP amplitudes, topography, and tomography [variable resolution electromagnetic tomography (VARETA)]. The results revealed that SSVEPs related to background processing showed higher amplitudes in the consistent as opposed to the inconsistent condition, whereas object-related SSVEPs showed the reversed pattern of effects. Given the SSVEPs' sensitivity to visual attention, the results indicate that semantic inconsistency leads to greater attention focused on the object. If all image parts are semantically related, attention is rather directed to the background. The attentional advantage to inconsistent objects in a scene is likely the result of a mismatch between background-based expectations and semantic object information. A clear lateralization of the consistency effect in the anterior temporal lobes indicates functional hemispheric asymmetries in processing background- and object-related semantic information. In summary, the present study is the first to demonstrate the feasibility of SSVEPs to unravel the respective contributions of concurrent neuronal processes involved in the perception of background and object.

Electrophysiological (EEG/MEG) imaging challenges statistics by providing two views of the same underlying spatio‐temporal brain activity: a topographic view (EEG/MEG) and tomographic view (EEG/MEG source reconstructions). It is a common practice that statistical parametric mapping (SPM) for these two situations is developed separately. In particular, assessing statistical significance of functional connectivity is a major challenge in these types of studies. This work introduces statistical tests for assessing simultaneously the significance of spatio‐temporal correlation structure between ERP/ERF components as well as that of their generating sources. We introduce a greatest root statistic as the multivariate test statistic for detecting functional connectivity between two sets of EEG/MEG measurements at a given time instant. We use some new results in random field theory to solve the multiple comparisons problem resulting from the correlated test statistics at each time instant. In general, our approach using the union‐intersection (UI) principle provides a framework for hypothesis testing about any linear combination of sensor data, which allows the analysis of the correlation structure of both topographic and tomographic views. The performance of the proposed method is illustrated with real ERP data obtained from a face recognition experiment. Hum Brain Mapp 2009. © 2009 Wiley‐Liss, Inc.

The goal of this paper is to assess the P-value of local maxima of time-varying cross-correlation random fields. The motivation for this comes from an electroencephalography (EEG) experiment, where one seeks connectivity between all pairs of voxels inside the brain at each time point of the recording window. In this way, we extend the results of [Cao, J., Worsley, K.J., 1999b. The geometry of correlation fields with an application to functional connectivity of the brain. The Annals of Applied Probability 9 (4), 1021-1057] by searching for high correlations not only over all pairs of voxels, but over all time points as well. We apply our results to an EEG data set of a face recognition paradigm. Our analysis determines those time instants for which there are significantly correlated regions involved in face recognition.

One of the most interesting problems studied in Random Field Theory (RFT) is to approximate the distribution of the maximum of a random field. This problem usually appears in a general hypothesis testing framework, where the statistics of interest are the maximum of a random field of a known distribution. In this paper, we use the RFT approach to compare two independent correlation random fields, R 1 and R 2 . Our statistics of interest are the maximum of a random field G , resulting from the difference between the Fisher’s Z transformation of R 1 and R 2 , respectively. The Fisher’s Z transformation guarantees a Gaussian distribution at each point of G but, unfortunately, G is not transformed into a Gaussian random field. Hence, standard results of RFT for Gaussian random fields are not longer available for G . We show here that the distribution of the maximum of G can still be approximated by the distribution of the maximum of a Gaussian random field, provided there is some correction by its spatial smoothness. Indeed, we present a general setting to obtain this correction. This is done by allowing different smoothness parameters for the components of G . Finally, the performance of our method is illustrated by means of both numerical simulations and real Electroencephalography data, recorded during a face recognition experimental paradigm.

This paper extends a previously formulated deterministic metabolic/hemodynamic model for the generation of blood oxygenated level dependent (BOLD) responses to include both physiological and observation stochastic components (sMHM). This adds a degree of flexibility when fitting the model to actual data by accounting for un-modelled activity. We then show how the innovation method can be used to estimate unobserved metabolic/hemodynamic as well as vascular variables of the sMHM, from simulated and actual BOLD data. The proposed estimation method allowed for doing model comparison by calculating the model's AIC and BIC. This methodology was then used to select between different neurovascular coupling assumptions underlying sMHM. The proposed framework was first validated on simulations and then applied to BOLD data from a motor task experiment. The models under comparison in the analysis of the actual data considered that vascular response was coupled to: (I) inhibition, (II) excitation, (III) both excitation and inhibition. Data was best described by model II, although model III was also supported.

Natural stimuli are compiled of numerous features, which are cortically represented in dispersed structures. Synchronized oscillations in the Gamma-Band (> 30 Hz; induced Gamma-Band Responses, iGBRs), are regarded as a plausible mechanism to re-integrate these regions into a meaningful cortical object representation. Using electroencephalography (EEG) it was demonstrated that the generators of iGBRs can be localized to temporal, parietal, posterior, and frontal areas. The present magnetoencephalogram (MEG) study intended to replicate these findings in order contribute to the ongoing debate regarding the possible functional difference of high-frequency signals as measured by both techniques. During a standard object recognition task we found an augmentation of the iGBR after the presentation of meaningful as opposed to meaningless stimuli at ∼ 160–440 ms after stimulus onset. This peak was localized to inferior temporal gyri, superior parietal lobules and the right middle frontal gyrus. Importantly, most of these brain structures were significantly phase-locked to each other. The implications of these results are twofold: (1) they present further evidence for the view that iGBRs signify neuronal activity in a broadly distributed network during object recognition. (2) MEG is well suited to detect induced high-frequency oscillations with a very similar morphology as revealed by EEG recordings, thereby eliminating known problems with electroencephalographical methods (e.g. reference confounds). In contrast to the iGBR, the localization of event-related fields (ERFs) and evoked Gamma-Band Response (eGBRs) revealed generators in focal visual areas, and thus, seem to mirror early sensory processing.

This paper describes an application of hierarchical or empirical Bayes to the distributed source reconstruction problem in electro- and magnetoencephalography (EEG and MEG). The key contribution is the automatic selection of multiple cortical sources with compact spatial support that are specified in terms of empirical priors. This obviates the need to use priors with a specific form (e.g., smoothness or minimum norm) or with spatial structure (e.g., priors based on depth constraints or functional magnetic resonance imaging results). Furthermore, the inversion scheme allows for a sparse solution for distributed sources, of the sort enforced by equivalent current dipole (ECD) models. This means the approach automatically selects either a sparse or a distributed model, depending on the data. The scheme is compared with conventional applications of Bayesian solutions to quantify the improvement in performance.

This article proposes a Bayesian spatio-temporal model for source reconstruction of M/EEG data. The usual two-level probabilistic model implicit in most distributed source solutions is extended by adding a third level which describes the temporal evolution of neuronal current sources using time-domain General Linear Models (GLMs). These comprise a set of temporal basis functions which are used to describe event-related M/EEG responses. This places M/EEG analysis in a statistical framework that is very similar to that used for PET and fMRI. The experimental design can be coded in a design matrix, effects of interest characterized using contrasts and inferences made using posterior probability maps. Importantly, as is the case for single-subject fMRI analysis, trials are treated as fixed effects and the approach takes into account between-trial variance, allowing valid inferences to be made on single-subject data. The proposed probabilistic model is efficiently inverted by using the Variational Bayes framework under a convenient mean-field approximation (VB-GLM). The new method is tested with biophysically realistic simulated data and the results are compared to those obtained with traditional spatial approaches like the popular Low Resolution Electromagnetic TomogrAphy (LORETA) and minimum variance Beamformer. Finally, the VB-GLM approach is used to analyze an EEG data set from a face processing experiment.

We describe a Bayesian scheme to analyze images, which uses spatial priors encoded by a diffusion kernel, based on a weighted graph Laplacian. This provides a general framework to formulate a spatial model, whose parameters can be optimized. The application we have in mind is a spatiotemporal model for imaging data. We illustrate the method on a random effects analysis of fMRI contrast images from multiple subjects; this simplifies exposition of the model and enables a clear description of its salient features. Typically, imaging data are smoothed using a fixed Gaussian kernel as a pre-processing step before applying a mass-univariate statistical model (e.g., a general linear model) to provide images of parameter estimates. An alternative is to include smoothness in a multivariate statistical model (Penny, W.D., Trujillo-Barreto, N.J., Friston, K.J., 2005. Bayesian fMRI time series analysis with spatial priors. Neuroimage 24, 350–362). The advantage of the latter is that each parameter field is smoothed automatically, according to a measure of uncertainty, given the data. In this work, we investigate the use of diffusion kernels to encode spatial correlations among parameter estimates. Nonlinear diffusion has a long history in image processing; in particular, flows that depend on local image geometry (Romeny, B.M.T., 1994. Geometry-driven Diffusion in Computer Vision. Kluwer Academic Publishers) can be used as adaptive filters. This can furnish a non-stationary smoothing process that preserves features, which would otherwise be lost with a fixed Gaussian kernel. We describe a Bayesian framework that incorporates non-stationary, adaptive smoothing into a generative model to extract spatial features in parameter estimates. Critically, this means adaptive smoothing becomes an integral part of estimation and inference. We illustrate the method using synthetic and real fMRI data.

A biophysical model of the coupling between neuronal activity and the BOLD signal that allows for explicitly evaluating the role of both excitatory and inhibitory activity is formulated in this paper. The model is based on several physiological assumptions. Firstly, in addition to glycolysis, the “glycogen shunt” is assumed for excitatory synapses as a mechanism for energy production in the astrocytes. As a result, oxygen-to-glucose index (OGI) is not constant but varies with excitatory neuronal activity. In contrast, a constant OGI = 6 (glycolysis) is assumed for inhibitory synapses. Finally we assume that cerebral blood flow is not directly controlled by energy usage, but it is only related to excitatory activity. Simulations' results show that increases in excitatory activity amplify the oscillations associated with the transient BOLD response, by increasing the initial dip, the maximum, and the post-stimulus undershoot of the signal. In contrast, increasing the inhibitory activity evoked an overall decrease of the BOLD signal along the whole time interval of the response. Simultaneous increase of both types of activity is then expected to reinforce the initial dip and the post-stimulus undershoot, while respective effects on the maximum tend to counteract each other. Two mechanisms for negative BOLD response (NBS) generation were predicted by the model: (i) when inhibition was present alone or together with low activation levels and (ii) when deactivation occurred independently of the accompanying inhibition level. Interestingly, NBS was associated with negative oxygen consumption changes only for the case of mechanism (ii).

This note derives the variational free energy under the Laplace approximation, with a focus on accounting for additional model complexity induced by increasing the number of model parameters. This is relevant when using the free energy as an approximation to the log-evidence in Bayesian model averaging and selection. By setting restricted maximum likelihood (ReML) in the larger context of variational learning and expectation maximisation (EM), we show how the ReML objective function can be adjusted to provide an approximation to the log-evidence for a particular model. This means ReML can be used for model selection, specifically to select or compare models with different covariance components. This is useful in the context of hierarchical models because it enables a principled selection of priors that, under simple hyperpriors, can be used for automatic model selection and relevance determination (ARD). Deriving the ReML objective function, from basic variational principles, discloses the simple relationships among Variational Bayes, EM and ReML. Furthermore, we show that EM is formally identical to a full variational treatment when the precisions are linear in the hyperparameters. Finally, we also consider, briefly, dynamic models and how these inform the regularisation of free energy ascent schemes, like EM and ReML.

The neural basis of selective attention within hierarchically organized Navon figures has been extensively studied with event related potentials (ERPs), by contrasting responses obtained when attending the global and the local echelons. The findings are inherently ambiguous because both levels are always presented together. Thus, only a mixture of the brain responses to two levels can be observed. Here, we use a method that allows unveiling of global and local letters at distinct times, enabling estimation of separate ERPs related to each level. Two interspersed oddball streams were presented, each using letters from one level and comprised of frequent distracters and rare targets. Previous work and our Experiment 1 show that it is difficult to divide attention between two such streams of stimuli. ERP recording in Experiment 2 evinced an early selection negativity (SN, with latencies to the 50% area of about 266 ms for global distracters and 276 ms for local distracters) that was larger for attended relative to unattended distracters. The SN was larger over right posterior occipito-temporal derivations for global stimuli and over left posterior occipito-temporal derivations for local stimuli (although the latter was less strongly lateralized). A discrimination negativity (DN, accompanied by a P3b) was larger for attended targets relative to attended distracters, with latencies to the 50% area of about 316 ms for global stimuli and 301 ms for local stimuli, which presented a similar distribution for both levels over left temporo-parietal electrodes. The two negativities apparently index successive stages in the processing of a selected level within a compound figure. By resolving the ambiguity of traditional designs, our method allowed us to observe the effects of attention for each hierarchical level on its own.

This article summarizes the theoretical foundations of Bayesian model inference as regards model selection and model averaging. Although not formally treated here, model combination is also motivated as a variant of Bayesian model averaging. The main practical implementations and approximations to model inference are reviewed, and the relationships among the different approaches are highlighted.

The formation of cortical object representations requires the activation of cell assemblies, correlated by induced oscillatory bursts of activity > 20 Hz (induced gamma band responses; iGBRs). One marker of the functional dynamics within such cell assemblies is the suppression of iGBRs elicited by repeated stimuli. This effect is commonly interpreted as a signature of 'sharpening' processes within cell-assemblies, which are behaviourally mirrored in repetition priming effects. The present study investigates whether the sharpening of primed objects is an automatic consequence of repeated stimulus processing, or whether it depends on task demands. Participants performed either a 'living/non-living' or a 'bigger/smaller than a shoebox' classification on repeated pictures of everyday objects. We contrasted repetition-related iGBR effects after the same task was used for initial and repeated presentations (no-switch condition) with repetitions after a task-switch occurred (switch condition). Furthermore, we complemented iGBR analysis by examining other brain responses known to be modulated by repetition-related memory processes (evoked gamma oscillations and event-related potentials; ERPs). The results obtained for the 'no-switch' condition replicated previous findings of repetition suppression of iGBRs at 200-300 ms after stimulus onset. Source modelling showed that this effect was distributed over widespread cortical areas. By contrast, after a task-switch no iGBR suppression was found. We concluded that iGBRs reflect the sharpening of a cell assembly only within the same task. After a task switch the complete object representation is reactivated. The ERP (220-380 ms) revealed suppression effects independent of task demands in bilateral posterior areas and might indicate correlates of repetition priming in perceptual structures.

The formation of cortical object representations requires the activation of cell assemblies, correlated by induced oscillatory bursts above 20 Hz (gamma band), which are characterized by trial-by-trial latency fluctuations around a mean of approximately 300 ms after stimulus onset. The present electroencephalogram (EEG) study was intended to uncover to the generators of induced gamma band responses (GBRs) and to analyze phase-synchronization between these sources. A standard object recognition task was used to elicit gamma activity. At the scalp surface (electrode space), we found an augmentation of induced GBRs after the presentation of meaningful (familiar) as opposed to meaningless (unfamiliar) stimuli, which was accompanied by a dense pattern of significant phase-locking values between distant recording sites. Subsequently, intracranial current density distributions compatible with the observed scalp voltage topographies were estimated by means of VARETA (Variable Resolution Electromagnetic Tomography). In source space brain electrical tomographies (BETs) revealed widespread generators of induced GBRs at temporal, parietal, posterior, and frontal areas. Phase-locking analysis was calculated between re-constructed electrode signals based on separate forward solutions of the observed generators, thereby eliminating the possibly confounding influence of activity from areas not under observation. The results support the view that induced GBRs signify synchronous neuronal activity in a broadly distributed network during object recognition. The localization of the generators of event-related potentials (ERPs), evoked gamma activity, and induced alpha activity revealed different sources as compared to the induced GBR and, thus, seem to mirror complementary functions during the present task as compared to induced high-frequency brain dynamics.

We describe a Bayesian estimation and inference procedure for fMRI time series based on the use of General Linear Models (GLMs). Importantly, we use a spatial prior on regression coefficients which embodies our prior knowledge that evoked responses are spatially contiguous and locally homogeneous. Further, using a computationally efficient Variational Bayes framework, we are able to let the data determine the optimal amount of smoothing. We assume an arbitrary order Auto-Regressive (AR) model for the errors. Our model generalizes earlier work on voxel-wise estimation of GLM-AR models and inference in GLMs using Posterior Probability Maps (PPMs). Results are shown on simulated data and on data from an event-related fMRI experiment.

In this paper, the Bayesian Theory is used to formulate the Inverse Problem (IP) of the EEG/MEG. This formulation offers a comparison framework for the wide range of inverse methods available and allows us to address the problem of model uncertainty that arises when dealing with different solutions for a single data. In this case, each model is defined by the set of assumptions of the inverse method used, as well as by the functional dependence between the data and the Primary Current Density (PCD) inside the brain. The key point is that the Bayesian Theory not only provides for posterior estimates of the parameters of interest (the PCD) for a given model, but also gives the possibility of finding posterior expected utilities unconditional on the models assumed. In the present work, this is achieved by considering a third level of inference that has been systematically omitted by previous Bayesian formulations of the IP. This level is known as Bayesian model averaging (BMA). The new approach is illustrated in the case of considering different anatomical constraints for solving the IP of the EEG in the frequency domain. This methodology allows us to address two of the main problems that affect linear inverse solutions (LIS): (a) the existence of ghost sources and (b) the tendency to underestimate deep activity. Both simulated and real experimental data are used to demonstrate the capabilities of the BMA approach, and some of the results are compared with the solutions obtained using the popular low-resolution electromagnetic tomography (LORETA) and its anatomically constraint version (cLORETA).

Previous studies with compound letters have not been able to dissociate in time the presentation of the global and local levels. Therefore, direct timing of attentional shifts between levels has not been possible, nor has the Hillyard sustained-selective attention paradigm been fully applied. Here we describe work with a new approach, rapid serial object transformation, which shows that shifts of attention from the local to the global level are faster than in the opposite direction in typical observers. The method also enables independent estimates of the neural responses to local and global aspects, and was used to show that selection of one hierarchal level modulates early occipital–temporal components with a shorter latency effect for responses to global letters. The use of the method for basic research in visual perception is discussed and preliminary work with autistic observers (who have longer local to global than global to local attentional shifts) illustrates the potential of the paradigm for studying psychopathology.

The auditory processing of self-generated sounds is characterized by an attenuated vertex N1-component of the event-related potential (ERP) compared to the responses elicited by externally generated sounds. Typically, a motor condition where sounds are actively produced by button presses is compared with a passive listening condition. While this effect is usually interpreted as reflection of an internal forward model system, the impact of attention and arousal on the so called self-generation effect has not been systematically controlled in these studies: Is the auditory stimulation more attended during the active task compared to passive listening, e.g., caused by a higher arousal level? Or is it rather attended less and attention is drawn away from the task-irrelevant stimulation to the motor task? Accordingly, the self-generation effects reported in the literature can easily be over- or underestimated. In the present study we disentangled attention from the self-generation effect by introducing an active listening condition, in which attention is focused to the same feature as in the self-generation condition — the stimulus onset-to-onset interval. We observed a classical ‘self-generation effect’, i.e. attenuated amplitudes for self-generated compared to passive listened sounds at frontocentral electrodes. As expected this effect was overlapped by attention effects in space and time. However, topographical and tomographical analyses allowed us to clearly disentangle both effects. Our results argue for the existence of a genuine self-generation effect, but emphasize the problem of possible over- or underestimation caused by attentional confounds.

The lack of clear understanding of the pathophysiology of chronic pain could explain why we currently have only a few effective treatments. Understanding how pain relief is realised during placebo analgesia could help develop improved treatments for chronic pain. Here, we tested whether experimental placebo analgesia was associated with altered resting-state cortical activity in the alpha frequency band of the electroencephalogram (EEG). Alpha oscillations have been shown to be influenced by top-down processes, which are thought to underpin the placebo response. Seventy-three healthy volunteers, split into placebo or control groups, took part in a well-established experimental placebo procedure involving treatment with a sham analgesic cream. We recorded ongoing (resting) EEG activity before, during, and after the sham treatment. We show that resting alpha activity is modified by placebo analgesia. Post-treatment, alpha activity increased significantly in the placebo group only (p < 0.001). Source analysis suggested that this alpha activity might have been generated in medial components of the pain network, including dorsal anterior cingulate cortex, medial prefrontal cortex, and left insula. These changes are consistent with a cognitive state of pain expectancy, a key driver of the placebo analgesic response. The manipulation of alpha activity may therefore present an exciting avenue for the development of treatments that directly alter endogenous processes to better control pain.

Brain Electric/Magnetic Tomography (BET /BMT) consists of a 3D-image reconstruction of the Primary Current Density inside the brain from the signals measured outside the head, which can be considered as a functional neuroimaging modality. Mathematically, this is known as the EEG/MEG inverse problem (IP), which is ill posed (non-unique solution). To find a unique BET/BMT, additional in-formation and proper models are needed, which has led to the development of several different methods for solving the IP.

We introduce a new generative model of the Encephalography (EEG/MEG) data, the inversion of which allows for inferring the locations and temporal evolution of the underlying sources as well as their dynamical interactions. The proposed Switching Mesostate Space Model (SMSM) builds on the multi-scale generative model for EEG/MEG by Daunizeau and Friston (2007). SMSM inherits the assumptions that (1) bioelectromagnetic activity is generated by a set of distributed sources, (2) the dynamics of these sources can be modelled as random fluctuations about a small number of mesostates, and (3) the number of mesostates engaged by a cognitive task is small. Additionally, four generalising assumptions are now included: (4) the mesostates interact according to a full Dynamical Causal Network (DCN) that can be estimated; (5) the dynamics of the mesostates can switch between multiple approximately linear operating regimes; (6) each operating regime remains stable over finite periods of time (temporal clusters); and (7) the total number of times the mesostates' dynamics can switch is small. The proposed model adds, therefore, a level of flexibility by accommodating complex brain processes that cannot be characterised by purely linear and stationary Gaussian dynamics. Importantly, the SMSM furnishes a new interpretation of the EEG/MEG data in which the source activity may have multiple discrete modes of behaviour, each with approximately linear dynamics. This is modelled by assuming that the connection strengths of the underlying mesoscopic DCN are time-dependent but piecewise constant, i.e. they can undergo discrete changes over time. A Variational Bayes inversion scheme is derived to estimate all the parameters of the model by maximising a (Negative Free Energy) lower bound on the model evidence. This bound is used to select among different model choices that are defined by the number of mesostates as well as by the number of stationary linear regimes. The full model is compared to a simplified version that uses no dynamical assumptions as well as to a standard EEG inversion technique. The comparison is carried out using an extensive set of simulations, and the application of SMSM to a real data set is also demonstrated. Our results show that for experimental situations in which we have some a priori belief that there are multiple approximately linear dynamical regimes, the proposed SMSM provides a natural modelling tool.

Does temporal regularity facilitate prediction in audition? To test this, we recorded human event-related potentials to frequent standard tones and infrequent pitch deviant tones, pre-attentively delivered within isochronous and anisochronous (20% onset jitter) rapid sequences. Deviant tones were repeated, either with high or low probability. Standard tone repetition sets a first-order prediction, which is violated by deviant tone onset, leading to a first-order prediction error response (Mismatch Negativity). The response to highly probable deviant repetitions is, however, attenuated relative to less probable repetitions, reflecting the formation of higher-order sensory predictions. Results show that temporal regularity is required for higher-order predictions, but does not modulate first-order prediction error responses. Inverse solution analyses (Variable Resolution Electrical Tomography; VARETA) localized the error response attenuation to posterior regions of the left superior temporal gyrus. In a control experiment with a slower stimulus rate, we found no evidence for higher-order predictions, and again no effect of temporal information on first-order prediction error. We conclude that: (i) temporal regularity facilitates the establishing of higher-order sensory predictions, i.e. 'knowing what next', in fast auditory sequences; (ii) first-order prediction error relies predominantly on stimulus feature mismatch, reflecting the adaptive fit of fast deviance detection processes.