This talk presents a few studies that aim to interpret oscillatory travelling waves in the predictive coding framework. In the first part, I’ll introduce a simple model of the visual cortex based on predictive coding mechanisms, in which physiological communication delays between levels generate alpha-band rhythms. Interestingly, these oscillations propagate as traveling waves across levels, both forward (during visual stimulation) and backward (during rest). Remarkably, experimental EEG data matched the predictions of our model. In the second part of the talk, I’ll present two studies that indirectly investigate the link between predictive coding mechanisms and traveling waves experimentally: the first one investigates the effect of a powerful psychedelics drug, N,N, dimethyltryptamine (DMT), on alpha-band oscillations, and the second one interprets the pattern of oscillatory traveling waves in schizophrenic patients in the light of Predictive Coding. In the last part of the talk, I will show some preliminary results on a statistical learning paradigm that directly explores the link between traveling waves and predictive coding processes.

Electroencephalography (EEG) is widely utilized in clinical settings due to its cost-effectiveness, portability, and non-invasive nature, making it one of the primary methods for directly measuring neuronal activity. In epilepsy, it plays a crucial role in its diagnosis, as well as in providing prognostic assessments and follow-up examinations. Epilepsy is nowadays considered a brain network disease and the analyses of EEG-derived networks is showing its potential for clinical applications. In this presentation, we will cover recent findings on functional connectivity and graph analyses and how, when applied to epilepsy, can help 1) the diagnosis in absence of interictal discharges and 2) the prediction of surgical outcome. Finally, we will focus on a novel approach called graph signal processing (GSP), which integrates electrophysiology with information on brain structure extracted from diffusion imaging. We will uncover the methodological steps involved in GSP and discuss the potential of a surgical biomarker based on structure-function connectivity.

The environment in which we live is in a state of constant change. This places adaptive demands on the sensorimotor system, which is required to constantly integrate multimodal information to generate estimates of body position and state in space. Thus, sensorimotor adaptation is involved in the creation of internal models and the development of one's own body representation, which are updated throughout life to enable smooth interactions with the environment. Inefficient integration of visual and proprioceptive information can lead to postural deficit which can be assessed on a behavioral level. In addition, electroencephalography (EEG) provides complementary insight into the involvement of the central nervous system in sensorimotor adaptation processes. In this talk, we will present preliminary results from two longitudinal studies from our laboratory, focusing on the effects of rehabilitation programs for adolescents and older adults on sensorimotor integration and its EEG and behavioral correlates. Our results highlight the involvement of specific brain oscillations, particularly the alpha band, as biomarkers of postural demand and adaptability.

EEG studies in patients with schizophrenia typically analyze the data using a single method. The result often shows a significant difference between patients and healthy controls. Subsequent studies then try to describe the underlying neural and genetic abnormal mechanism. This approach rests on the implicit assumption that the analysis being used taps into a key abnormality of schizophrenia. If true, different analysis methods showing group differences should strongly correlate with each other. Here, we explicitly tested this assumption using resting-state EEG from 121 patients with schizophrenia and 75 healthy controls. We extracted 194 EEG features, including EEG microstates, source and sensor connectivity, and spectral features, some of which have long been investigated in schizophrenia research. We found that 69 of the 194 EEG features showed significant group differences between patients and healthy controls with medium to large effect sizes (Cohen d ranged from 0.46 to 1.04), indicating that important mechanisms are targeted. Surprisingly, the correlations between these 69 EEG features were generally low in the patient population, suggesting that different EEG features target different mechanisms of schizophrenia. As such, they may be less representative of schizophrenia than initially thought. Importantly, we show that most of these EEG features were remarkably stable even after several years, i.e., test-retests were high, indicating that there is a large variability in the patient population that cannot be summarized by one method. We propose that employing a range of analysis methods and outcomes from multiple paradigms may help to identify key mechanisms of schizophrenia.

Neuroscientists, classically treat visual evoked potential (VEP) inter-individual variability as a nuisance and eliminate it by aggregating individuals (grand-group average) or by reducing the waveforms to few summarizing features (peak amplitude and latency). As a consequence, the conclusions made on group-level canonical waveforms do not always generalize well to individuals. To illustrate this, let us consider visual backward masking (VBM). At the group-level, this well-established paradigm elicits in the visual cortex a late N1 component at around 200 ms after the onset of the target stimulus and is used to study sensory processing and selective attention. We used global field power (GFP) to summarize the global activity over all recording electrodes at each time point and found highly stable group average GFP waveforms and N1 components after 5 and 10 years using a longitudinal cohort of schizophrenia patients and healthy controls. However, we also observed a remarkable diversity of GFP waveforms across participants, with a majority exhibiting substantial deviations from the group-level average. For example, we noticed important variations in the number of peaks, latencies and amplitudes of N1 and interestingly, those variations remain very stable after 5 and 10 years. The high intra-individual stability suggests that the inter-individual differences in VEP contain meaningful information about the subjects which could notably be explained by variations in anatomy and/or cognition.

Alpha oscillations, which represent the brain's electrical activity in the frequency range of 8 to 13 Hz, are often thought to reflect the rhythm of consciousness. For instance, it has been suggested that alpha oscillations sample visual information into discrete conscious percepts at each cycle, and that their frequency determines the temporal resolution of visual perception. However, while studies on perceptual resolution provide information about what is consciously perceived, they cannot conclusively demonstrate when a discrete conscious percept emerges. Moreover, the cycles of alpha oscillations, lasting about 100 milliseconds, cannot explain the existence of windows of integration extending over several hundred milliseconds. In this talk, I will question the link between alpha oscillations and rhythmic sampling by demonstrating long-lasting effects of alpha oscillations in visual processing, which can only be attributed to the modulation of the content of conscious perception beyond the duration of a single alpha cycle. Taken together, these findings offer new evidence of the functional implications of alpha oscillations in visual perception over longer timescales than previously suggested.

Working memory (WM) is the brain’s ability to temporarily hold and manipulate information when it is no longer present in our environment. Previous studies showed that WM is supported by interactions between frontal-parietal areas and posterior sensory regions. In particular, the lateral prefrontal cortex (LPFC) provides top-down control over sensory WM representations. The intraparietal sulcus (IPS) has also been shown to be involved in maintaining WM information. The neural mechanisms behind such frontal-parietal control remain underspecified. Different brain oscillations may establish the inter-areal communication within these cognitive networks. In this talk, I will show findings from my recent studies using neuroimaging and computational modeling, as well as TMS. The results obtained from these studies provide converging evidence for a dual role of alpha oscillations: (i) the suppression of irrelevant information following retro-cue and (ii) the maintenance of memoranda through phase-coding mechanism.

Recollecting the progression of events is relevant not only for episodic memory but also for envisioning future behaviors based on experience. Previous research in freely-moving rodents proposed a central role of the medial temporal lobe’s (MTL) involvement in spatial navigation and the formation of novel episodic memories. Here, we study whether real-world navigation and episodic memory paradigms elicit functional similarities in the human MTL, which might generalize to imagined navigation. Taking benefit of recent neurotechnological developments, we examined intracranial electroencephalography (iEEG) recorded from a chronically implanted device in five freely-moving humans receiving responsive neurostimulation therapy. Participants learned to navigate two routes, including four turns each, in an indoor room (14.6 × 13.5 m2) fully equipped with motion capture. After each real-world walk, participants walked on a treadmill while imagining walking these routes in their minds. In agreement with previous reports, transient theta oscillations were evident in each participant at spectral peaks in the 4-10 Hz frequency range. Short-lasting theta bouts frequently occurred prior to upcoming turns and formed temporal dynamics that encode the transitions between linear route segments consistently across trials (30-35, left/right walks each). During imagined navigation, we found that theta bouts occurring at specific time points resembled the routes’ geometry, similar to the real-world dynamics. This resemblance was absent during sole treadmill walking that we used as a control condition. We subsequently used these theta dynamics as priors to decompose simultaneously recorded scalp EEG and to suppress unrelated activity, such as motion artifacts. We identified a subspace of components that significantly co-modulated with the intracranial data. Using EEG source imaging, we found that the generators of specific components in every participant overlap with the actual location of the intracranial theta dynamics given by the implantation sites. In addition, we found other components that were functionally coupled to MTL dynamics, but their sources were localized to the prefrontal cortex. These findings open up novel possibilities for studying human MTL interactions with cortical areas in freely behaving and imagining humans. Altogether, our results open novel avenues for studying real-world spatial navigation, episodic memory, and imaginable future behaviors.