Milan Brázdil

Dynamic modeling of neuronal responses in fMRI using cubature Kalman filtering

This paper presents a new approach to inverting (fitting) models of coupled dynamical systems based on state-of-the-art (cubature) Kalman filtering. Crucially, this inversion furnishes posterior estimates of both the hidden states and parameters of a system, including any unknown exogenous input. Because the underlying generative model is formulated in continuous time (with a discrete observation process) it can be applied to a wide variety of models specified with either ordinary or stochastic differential equations. These are an important class of models that are particularly appropriate for biological time-series, where the underlying system is specified in terms of kinetics or dynamics (i.e., dynamic causal models). We provide comparative evaluations with generalized Bayesian filtering (dynamic expectation maximization) and demonstrate marked improvements in accuracy and computational efficiency. We compare the schemes using a series of difficult (nonlinear) toy examples and conclude with a special focus on hemodynamic models of evoked brain responses in fMRI. Our scheme promises to provide a significant advance in characterizing the functional architectures of distributed neuronal systems, even in the absence of known exogenous (experimental) input; e.g., resting state fMRI studies and spontaneous fluctuations in electrophysiological studies. Importantly, unlike current Bayesian filters (e.g. DEM), our scheme provides estimates of time-varying parameters, which we will exploit in future work on the adaptation and enabling of connections in the brain.

Error processing – evidence from intracerebral ERP recordings

Over the last decade, several authors have described an early negative (Ne) and a later positive (Pe) potential in scalp event-related potentials (ERPs) of incorrect choice reactions. The aim of the present study was to investigate the intracerebral origin and distribution of these potentials. Seven intractable epileptic patients participated in the study. A total of 231 sites in the frontal, temporal, and parietal lobes were investigated by means of depth electrodes. A standard visual oddball paradigm was performed, and electroencephalogram (EEG) epochs with correct and incorrect motor reactions were averaged independently. Prominent, mostly biphasic, ERP complexes resembling scalp Ne/Pe potentials were consistently observed in several cortical locations after incorrect trials. The most consistent findings were obtained from mesiotemporal structures; in addition to P3-like activity found after correct responses, an Ne/Pe complex was generally detected after incorrect trials. The Pe had a longer latency than the P3. Other generators of Ne/Pe-like potentials were located in different regions of the frontal lobe. The latency of the Ne was shortest in parietal, longer in temporal, and longest in frontal regions. Our findings firstly show that multiple cortical structures generate Ne and Pe. In addition to the rostral anterior cingulate cortex, the mesiotemporal and some prefrontal cortical sites seem to represent integral components of the brain’s errorchecking system. Secondly, the coupling of Ne and Pe to a complex suggests a common origin of Ne and Pe. Thirdly, the latency differences of the Ne across lobes suggest that the Ne is primarily elicited in posterior and temporal, and only later in frontal regions.

Intracerebral event-related potentials to subthreshold target stimuli.

OBJECTIVES Event-related potentials (ERPs) elicited by subthreshold visual stimuli were recorded directly from human frontal and temporal lobe structures to study unconscious perception. METHODS Thirteen intractable epileptic patients undergoing depth electrode recordings prior to their surgical treatment participated in the study. An original method of modified visual oddball paradigm with supraliminal and subliminal stimuli was applied, and the averaged responses to both kinds of stimuli were subsequently compared. RESULTS The results clearly prove that, at least from an electrophysiological viewpoint, the mechanism of unaware processing of visual stimuli in the human brain does not differ substantially from the aware processing. Finding the subliminal P3 waveform in a number of cortical structures (hippocampus and parahippocampal gyrus bilaterally, and left-sided mesiofrontal, orbitofrontal and lateral temporal cortex) indicates their involvement in unconscious processing, in spite of the fact that typical large-scale neurocognitive networks are not completely activated. The absence of activation consistently observed bilaterally in dorsolateral prefrontal cortices, in connection with right-sided cortical frontal lobe structures and right-sided lateral temporal neocortex in unconscious perception, supports the importance of these structures for the awareness of visual stimuli. The proof of the significantly faster unaware information processing represents another distinctive feature of implicit visual perception. CONCLUSIONS Based on the presented findings and comparisons with the results of previous ERP, functional magnetic resonance imaging, positron emission tomography, and clinical neuropsychological studies, a crucial role of the large-scale neural system for conscious experience of perception is suggested, which is distributed extensively among the dorsal posterior association areas and the prefrontal cortex, with the dominant part being that of the right hemisphere.

Dynamic Granger causality based on Kalman filter for evaluation of functional network connectivity in fMRI data

Increasing interest in understanding dynamic interactions of brain neural networks leads to formulation of sophisticated connectivity analysis methods. Recent studies have applied Granger causality based on standard multivariate autoregressive (MAR) modeling to assess the brain connectivity. Nevertheless, one important flaw of this commonly proposed method is that it requires the analyzed time series to be stationary, whereas such assumption is mostly violated due to the weakly nonstationary nature of functional magnetic resonance imaging (fMRI) time series. Therefore, we propose an approach to dynamic Granger causality in the frequency domain for evaluating functional network connectivity in fMRI data. The effectiveness and robustness of the dynamic approach was significantly improved by combining a forward and backward Kalman filter that improved estimates compared to the standard time-invariant MAR modeling. In our method, the functional networks were first detected by independent component analysis (ICA), a computational method for separating a multivariate signal into maximally independent components. Then the measure of Granger causality was evaluated using generalized partial directed coherence that is suitable for bivariate as well as multivariate data. Moreover, this metric provides identification of causal relation in frequency domain, which allows one to distinguish the frequency components related to the experimental paradigm. The procedure of evaluating Granger causality via dynamic MAR was demonstrated on simulated time series as well as on two sets of group fMRI data collected during an auditory sensorimotor (SM) or auditory oddball discrimination (AOD) tasks. Finally, a comparison with the results obtained from a standard time-invariant MAR model was provided.

Combined event-related fMRI and intracerebral ERP study of an auditory oddball task.

Event-related fMRI (efMRI) has been repeatedly used to seek the neural sources of endogenous event-related potentials (ERP). However, significant discrepancies exist between the efMRI data and the results of previously published intracranial ERP studies of oddball task. To evaluate the capacity of efMRI to define the sources of the P3 component of ERP within the human brain, both efMRI and intracerebral ERP recordings were performed in eight patients with intractable epilepsy (five males and three females) during their preoperative invasive video-EEG monitoring. An identical auditory oddball task with frequent and target stimuli was completed in two sessions. A total of 606 intracerebral sites were electrophysiologically investigated by means of depth electrodes. In accordance with the finding of multiple intracerebral generators of P3 potential, the target stimuli evoked MRI signal increase in multiple brain regions. However, regions with evident hemodynamic and electrophysiological responses overlapped only partially. P3 generators were always found within hemodynamic-active sites, if these sites were investigated by means of depth electrodes. On the other hand, unequivocal local sources of P3 potential were apparently also located outside the regions with a significant hemodynamic response (typically in mesiotemporal regions). Both methods should thus be viewed as mutually complementary in investigations of the spatial distribution of cortical and subcortical activation during oddball task.

Intracerebral Error-Related Negativity in a Simple Go/NoGo Task

Abstract: Performance monitoring represents a critical executive function of the human brain. In an effort to identify its anatomical and physiological aspects, a negative component of event-related potentials (ERPs), which occurs only on incorrect trials, has been used in the extensive investigation of error processing. This component has been termed “error-negativity” (Ne) or error-related negativity (ERN) and has been interpreted as a correlate of error detection. The aim of the present intracerebral ERP study was to contribute knowledge of the sources of the Ne/ERN, with a particular focus on the involvement of a frontomedian wall (FMW) in the genesis of this negativity. Seven patients with intractable epilepsy participated in the study. Depth electrodes were implanted to localize the seizure origin prior to surgical treatment. A total of 574 sites in the frontal, temporal, and parietal lobes were investigated. A simple Go/NoGo task was performed and EEG epochs with correct and erroneous motor respons...

Atypical hemispheric language dominance in left temporal lobe epilepsy as a result of the reorganization of language functions

Hemispheric language dominance, as determined by intracarotid short-acting barbiturate injections (Wada testing), was retrospectively evaluated in 44 right-handed patients with medically intractable left (n=26) or right (n=18) temporal lobe epilepsy (TLE). Atypical hemispheric language dominance (right or bilateral) was revealed in 13.6% of all patients investigated. A significantly higher rate of deviance from complete left hemisphere dominance was observed in left TLE (23.1%) than in right TLE (0%). Patient age at the time of seizure onset was proven a critical variable affecting atypical speech development in left TLE. The mean patient age at the time of seizure onset in left TLE patients with atypical language representation was 5.61 years, but was 13.13 years in patients with left hemisphere language dominance. Our results strongly support the influential impact of epileptiform activity spreading from the left temporal lobe structures on the reorganization of language functions in patients with early brain injury.

Reorganization of language‐related neuronal networks in patients with left temporal lobe epilepsy – an fMRI study

To investigate the inter‐ and intrahemispheric reorganization of the language cortex in left temporal lobe epilepsy (TLE) with left‐sided hippocampal sclerosis. A functional magnetic resonance imaging was performed on 13 right‐handed patients suffering from medically intractable left TLE, and in 13 sex‐ and age‐matched healthy controls. The activation paradigm used was a silent word generation task. A language laterality index (LI) was calculated from the number of activated voxels in the right and left anterior two‐thirds of the hemispheres. Significant differences between the patients and the controls were observed in the activation of the left‐sided inferior frontal gyrus. Less consistent findings in this region, as well as the relative protection of Brocas area from the activation, were revealed in the patients. In addition, different patterns of activation were proven in the cerebellum and other cortical as well as subcortical brain structures within both hemispheres. Significant differences were also found in the values of the language LIs between the investigated groups: these values suggested a more bihemispheric language representation in the patients. As anticipated, lateralization of the language functions in the epileptics significantly decreased in connection with an earlier age of initial insult. Our results support the hypothesis of a significant intra‐ and interhemispheric functional reorganization of language‐related neuronal networks in left TLE.

A SEEG study of ERP in motor and premotor cortices and in the basal ganglia.

OBJECTIVE Our intention was to study the electrical activity related to the cognitive processing of simple sensory stimuli in the brain structures that participate in motor control. We focused our interest on the 250-600 ms time window, in which cognitive activity most probably provides the basis for the activity recorded. METHODS Intracerebral stereoelectroencephalography (SEEG) recordings were made from 15 epilepsy surgery candidates. We studied potentials that were recorded in a time window in which P300 usually could be recorded on the scalp and that were directly recorded from brain structures involved in motor control: the primary motor cortex (MC, Brodmanns area 4); the lateral and mesial (SMA) premotor cortices (Brodmanns area 6); and the basal ganglia. We evaluated the first distinctive potential to occur in the 250-600 ms time window that displayed an amplitude gradient in several adjacent contacts. Four protocols were performed: an auditory oddball (aP3); a visual oddball (vP3); and contingent negative variation (CNV) protocols, in which the potentials evoked by the auditory warning (aCNV) and visual imperative (vCNV) stimuli were evaluated. In the protocols aP3, vP3, and vCNV, the tested person responded by flexing his/her thumb or hand. In the aCNV paradigm, and in a further auditory oddball paradigm (aP3c), no motor response was required. We compared the presence of an event-related potential (ERP) with an amplitude gradient to the absence of a generator. RESULTS The frequency of P3-like potential components was statistically significantly higher in the basal ganglia when compared with the explored cortical sites. Statistically non-significant latency differences between the basal ganglia and the cortex were displayed. The differences in the distribution of the potentials in the individual cortical areas were insignificant. The mean latency of vP3 was longer than the latencies of aP3, aP3c and vCNV. There was no significant difference between the distribution and latency of aP3 and aP3c. CONCLUSIONS (1) ERPs are generated in cortical as well as in subcortical structures. (2) The cognitive processing of sensory information in all the tested protocols occurred in the basal ganglia; the occurrence in the investigated cortical areas was less frequent and more dependent on the task. The basal ganglia may play an integrative role in cognitive information processing, in motor and non-motor tasks.

The role of frontal and temporal lobes in visual discrimination task — depth ERP studies

Visual event-related potentials were simultaneously recorded from different anatomical structures within frontal and temporal lobes in 12 epileptic patients. A simple discrimination task was performed to complement previous studies on the localization of P3 generators in the human brain. The role of multiple cortical structures in the generation of both P3a and P3b components was confirmed. Activities contemporary to a visual P3b were recorded in the hippocampus, amygdala and temporal pole. Anterior cingulate and orbitofrontal cortices-generated activities more closely related in time to the surface P3a. Earlier events related to visual discrimination took place in more lateral sites of the frontal lobe, but their contribution to the scalp P3 remains uncertain. Subsequently, mutual temporal relations among single generators were analyzed. The results suggested a processing-level hierarchy within the neural network for directed attention with a key role played by the dorsolateral prefrontal cortex.