The objectives of my PhD were to identify the spatial and the temporal dynamics of the brain areas involved in tactile change detection as well as the neural mechanisms responsible for the processing of tactile change detection. To that aim, three specific MEG studies were performed; each of them is addressing specific research aims.The first study investigated the spatiotemporal dynamics of the multilevel cortical processing of tactile change detection in human healthy subjects. This study disclosed a hierarchical organization from unimodal early tactile change detection at secondary somatosensory cortex to multi modal complex processing at bilateral temporo-parietal junctions, posterior parietal cortex and supplementary motor areas. The second study aimed at discriminating between debated neural mechanisms responsible for the genesis of the somatosensory mismatch negativity (sMMN). To do so, we manipulated the predictability of the deviant stimuli and the response to omissions in different kind of oddballs, the response to deviant stimuli paired with standards and occurring alone. We found out that mechanisms for early tactile change detection reflected by the sMMN were better explained by the predictive coding theory compared to the adaptation and adjustment theories. Finally we sought to characterize the alterations in early cortical tactile change detection in Friedreich Ataxia (FRDA); a neurological disorder characterized by somatosensory and cerebellar pathways degeneration. The aim of this work was to study the role of the cerebellum in the genesis of sMMN and its potential selectivity for somatosensory change detection compared to auditory. This study demonstrated that, in FRDA, both tactile and auditory pathways are affected at the level of primary sensory neurons and dorsal root/spiral ganglia in a genetically determined. By contrasts, early cortical sensory change detection in FRDA was impaired only in the tactile modality in line with the sMMN impairment described in patients with acquired cerebellar lesions or during cerebellar inhibition by trans cranial magnetic stimulation. These data brought novel empirical evidence supporting the contribution of spinocerebellar tracts in sMMN genesis at cSII cortex.In conclusion, this PhD contributed to identify the network responsible for tactile change detection that involves cuneocerebellar spinocerebellar tract and cSII cortex as somatosensory specific areas and TPJ, SMA & PPC as multimodal brain areas. We further provided evidence that early change detection mechanisms at SII cortex fall under the predictive coding framework and that change detection is hierarchically organized with inputs from low level areas for genesis of an adequate generative model of our environment and conscious representation of our body. / Doctorat en Sciences médicales (Médecine) / info:eu-repo/semantics/nonPublished

Dans cette thèse, nous avons étudié le codage prédictif and sa relation avec la perception et les oscillations. Nous avons, dans l'introduction, fait une revue des connaissances sur les neurones et le néocortex et un état de l'art du codage prédictif. Dans les chapitres principaux, nous avons tout d'abord, proposé l'idée, au travers d'une étude théorique, que la temporalité de la décharge crée une inhibition sélective dans les réseaux excitateurs non-sélectifs rétroactifs. Ensuite, nous avons montré les effets perceptuels du codage prédictif: la perception de la forme améliore la perception du contraste. Enfin, nous avons montré que le codage prédictif peut utiliser des oscillations dans différentes bandes de fréquences pour transmettre les informations en avant et en rétroaction. Cette thèse a fourni un mécanisme neuronal viable et innovant pour le codage prédictif soutenu par des données empiriques démontrant des prédictions rétroactives excitatrices et une relation forte entre codage prédictif et oscillations. / In this thesis, we investigated predictive coding and its relationship with perception and oscillations. We first reviewed my current understanding about facts of neuron and neocortex and state-of-the-arts of predictive coding in the introduction. In the main chapters, firstly, we proposed the idea that correlated spike times create selective inhibition in a nonselective excitatory feedback network in a theoretical study. Then, we showed the perceptual effect of predictive coding: shape perception enhances perceived contrast. At last, we showed that predictive coding can use oscillations with different frequencies for feedforward and feedback. This thesis provided an innovative and viable neuronal mechanism for predictive coding and empirical evidence for excitatory predictive feedback and the close relationship between the predictive coding and oscillations.

The mammalian visual system is an incredibly complex computation device, capable of performing the various tasks of seeing: navigation, pattern and object recognition, motor coordination, trajectory extrapolation, among others. Decades of research has shown that experience-dependent plasticity of cortical circuitry underlies the impressive ability to rapidly learn many of these tasks and to adjust as required. One particular thread of investigation has focused on unsupervised learning, wherein changes to the visual environment lead to corresponding changes in cortical circuits. The most prominent example of unsupervised learning is ocular dominance plasticity, caused by visual deprivation to one eye and leading to a dramatic re-wiring of cortex. Other examples tend to make more subtle changes to the visual environment through passive exposure to novel visual stimuli. Here, we use one such unsupervised paradigm, sequence learning, to study experience-dependent plasticity in the mouse visual system. Through a combination of theory and experiment, we argue that the mammalian visual system is an unsupervised learning device. Beginning with a mathematical exploration of unsupervised learning in biology, engineering, and machine learning, we seek a more precise expression of our fundamental hypothesis. We draw connections between information theory, efficient coding, and common unsupervised learning algorithms such as Hebbian plasticity and principal component analysis. Efficient coding suggests a simple rule for transmitting information in the nervous system: use more spikes to encode unexpected information, and fewer spikes to encode expected information. Therefore, expectation violations ought to produce prediction errors, or brief periods of heightened firing when an unexpected event occurs. Meanwhile, modern unsupervised learning algorithms show how such expectations can be learned. Next, we review data from decades of visual neuroscience research, highlighting the computational principles and synaptic plasticity processes that support biological learning and seeing. By tracking the flow of visual information from the retina to thalamus and primary visual cortex, we discuss how the principle of efficient coding is evident in neural activity. One common example is predictive coding in the retina, where ganglion cells with canonical center-surround receptive fields compute a prediction error, sending spikes to the central nervous system only in response to locally-unpredictable visual stimuli. This behavior can be learned through simple Hebbian plasticity mechanisms. Similar models explain much of the activity of neurons in primary visual cortex, but we also discuss ways in which the theory fails to capture the rich biological complexity. Finally, we present novel experimental results from physiological investigations of the mouse primary visual cortex. We trained mice by passively exposing them to complex spatiotemporal patterns of light: rapidly-flashed sequences of images. We find evidence that visual cortex learns these sequences in a manner consistent with efficient coding, such that unexpected stimuli tend to elicit more firing than expected ones. Overall, we observe dramatic changes in evoked neural activity across days of passive exposure. Neural responses to the first, unexpected sequence element increase with days of training while responses at other, expected time points either decrease or stay the same. Furthermore, substituting an unexpected element for an expected one or omitting an expected element both cause brief bursts of increased firing. Our results therefore provide evidence for unsupervised learning and efficient coding in the mouse visual system, especially because unexpected events drive prediction errors. Overall, our analysis suggests novel experiments, which could be performed in the near future, and provides a useful framework to understand visual perception and learning.

This dissertation concerns the alteration of functional connectivity in the human brain through different patterns of somatosensory stimulation. In particular, I distinguish whether stimuli are regular (i.e., expected by the subject) or irregular (i.e., unexpected by the subject). An emerging theory of brain function known as Predictive Coding states that the brain is continuously creating an internal model of its environment that is constantly trying to predict what is going to happen. Expected sensory input leads to model consol- idation, while unexpected input leads to model update. In this context it is assumed that central neuronal processing differs significantly between these two cases. Furthermore, in my experiments, the stimulation is applied in two more variants which are also believed to be processed in completely different ways: consciously perceptible (suprathreshold) and imperceptible (subthreshold). To measure functional connectivity in the acquired fMRI data, a method referred to as eigenvector centrality mapping (ECM) was chosen. This method has gained increasing attention in the fMRI community, as it represents a whole-brain approach that can be ap- plied for resting-state experiments. While there are a number of other centrality measures, each with their advantages and disadvantages, ECM stands out as being parameter-free and does not depend on prior assumptions. Similar to Google’s Pagerank algorithm, it assigns areas (“nodes”) in a network with a high centrality score that are closely connected to other central areas as well. Generally, increased connectivity is interpreted as of greater “importance” to the network. As there are different approaches on how to calculate ECM, I critically examine these and delve deeper into the method itself. Three main research questions guided this study: 1. Is there a brain connectivity (ECM) alteration in the human brain for somatosensory stimulation that is pattern dependent (7 Hz irregular vs. regular)? 2. Is there a brain connectivity (ECM) alteration in the human brain for somatosensory stimulation that is intensity dependent (7 Hz suprathreshold vs. subthreshold)? 3. Are these different somatosensory stimulations (subthreshold, suprathreshold, irregular, regular) accompanied by a subsequent behavioral change? Two experiments were conducted. In Experiment 1, participants were exposed to all four stimulation conditions consecutively. Results showed significant ECM alterations compared to the initial baseline, suggesting persisting effects. To counter this, Experiment 2 adopted a different approach. Here, individual stimulations were applied to separate groups, with an additional control group for comparison. The results from Experiment 2 revealed that irregular stimulation compared to regular showed decreased connectivity in specific brain regions, aligning with the Predictive Coding theory. Suprathreshold stimulation showed increased connectivity in areas related to sensory input integration, possibly linked to conscious perception. Furthermore, all participants, regardless of stimulation type, showed heightened connectivity in somatosensory regions, suggesting a shared focus on tactile anticipation. The behavioral session from Experiment 2 found that irregular suprathreshold stimulation led to a decreased sensitivity to near-threshold stimuli. However, this change wasn't mirrored in the functional connectivity data. In conclusion, this research validated the differential neural processing of various somatosensory stimulations, supporting the Predictive Coding theory. The study also underscored the challenges and considerations in using ECM, particularly urging caution with methods that combine positive and negative correlations.:1 - Personal Motivation 2 - Introduction 3 - Background 3.1 Cognitive Neuroscience 3.2 Predictive Coding and the Free Energy Principle 3.3 Functional Magnetic Resonance Imaging (fMRI) 3.4 Blood-Oxygen Level Dependent (BOLD) Signal 3.5 Resting-State Functional Connectivity (RSFC) 3.6 Eigenvector Centrality Mapping (ECM) 3.7 Previous ECM Experiments - an Overview 3.8 Statistical Remarks 4 - Materials and Methods 4.1 Experimental Setup 4.1.1 Subjects 4.1.2 Experimental Procedures 4.1.2.1 Electrical stimulation 4.1.2.2 Absolute detection threshold examination 4.1.2.3 fMRI and behavioral data acquisition 4.2 fMRI Data Preprocessing 4.2.1 Prior Steps 4.2.2 Slice Time Correction 4.2.3 Motion Correction 4.2.4 Coregistration, Segmentation and Normalization 4.2.5 Spatial Filtering 4.2.6 Temporal Filtering 4.2.7 Grey Matter, White Matter and Cerebrospinal Fluid Masking 4.2.8 Nuisance Regression 4.3 ECM Approaches 4.4 Flexible Factorial Design 4.5 Seed-Based Functional Connectivity Analysis 5 - Results 5.1 Experiment 1 5.1.1 Detailed Results 5.1.2 Summary Experiment 1 5.2 Experiment 2 - fMRI Session 5.2.1 Detailed Results - ADD Approach 5.2.2 Detailed REsults - POS / NEG / ABS approach 5.2.3 Summary Experiment 2 5.3 ECM Approach Differences Illustrated by Examples 5.4 Seed-Based Functional Connectivity Analysis 5.5 Distribution of Voxel Timie Series Correlation Foefficients 5.6 Experiment 2 - Behavioral Session 6 - Discussion 6.1 Interpretation of Experimental Results 6.1.1 Experiment 1 6.1.2 Experimemnt 2 - fMRI Session 6.1.3 Experiment 2 - Behavioral Session 6.2 ECM Approaches 6.3 Further Considerations and Future Outlook 7 - Conclusion

<p>In many scientific, medical and defense applications of image/video compression, an <em>l</em>_∞error bound is required. However, pure <em>l</em>_∞-optimized image coding, colloquially known as near-lossless image coding, is prone to structured errors such as contours and speckles if the bit rate is not sufficiently high; moreover, previous <em>l</em>_∞-based image coding methods suffer from poor rate control. In contrast, the <em>l</em>₂ error metric aims for average fidelity and hence preserves the subtlety of smooth waveforms better than the <em>l</em>_∞ error metric and it offers fine granularity in rate control; but pure <em>l</em>₂-based image coding methods (e.g., JPEG 2000) cannot bound individual errors as <em>l</em>_∞-based methods can. This thesis presents a new compression approach to retain the benefits and circumvent the pitfalls of the two error metrics.</p> / Master of Applied Science (MASc)

The following study is part of a larger study examining techniques that might be of use in the release program of the Puerto Rican Parrot (Amazona vittata), including marking, capturing, and radio-tracking. The portion of the study reported here documents the vocal behavior of A. vittata during the reproductive season and examines the possibility of using vocalizations to identify individuals, determine the sex of individuals and determine the location of an individual's breeding territory. Objectives of this study included: 1) cataloguing and categorizing the vocal repertoire of A. vittata, 2) determining whether the vocal repertoire was sex-specific and region-specific and 3) determining if an individual's vocal repertoire could be used to identify it. The vocal repertoire was characterized using a hierarchical method and 147 calls were described. The repertoire was found contain a high percentage (76 %) of graded calls. Evolutionary strategies that may explain the complexity of such a repertoire are discussed. The vocal repertoire was found to be both sex- and region-specific. Characteristics analyzed included time and frequency parameters of sonagrams. Three methods were used to determine the feasibility of vocal recognition of individuals. These methods included: bird-call pairing, sonagraphic analysis, and linear predictive coding. Sonagraphic analyses in combination with linear predictive coding techniques show the most promise as tools in voice recognition of the parrot, however, further research will be necessary to determine how reliable voice recognition may be as a method for identifying individuals in the field. / Master of Science

No / Motion is known to distort visual space, producing illusory mislocalizations for flashed objects. Previously, it has been shown that when a stationary bar is flashed in the proximity of a moving stimulus, the position of the flashed bar appears to be shifted in the direction of nearby motion. A model consisting of predictive projections from the sub-system that processes motion information onto the sub-system that processes position information can explain this illusory position shift of a stationary flashed bar in the direction of motion. Based on this model of motion¿position interactions, we predict that the perceived position of a flashed stimulus should also be attracted towards a nearby moving stimulus. In the first experiment, observers judged the perceived vertical position of a flash with respect to two horizontally moving dots of unequal contrast. The results of this experiment were in agreement with our prediction of attraction towards the high contrast dot. We obtained similar findings when the moving dots were replaced by drifting gratings of unequal contrast. In control experiments, we found that neither attention nor eye movements can account for this illusion. We propose that the visual system uses predictive influences from the motion processing sub-system on the position processing sub-system to overcome the temporal limitations of the position processing system.

The human auditory system excels at detecting patterns necessary for processing speech and music. This system is adept at detecting changes to the incoming sounds. According to predictive coding theories, the brain generates hypothesis about what the incoming tone should be, and if the incoming tone does not match the hypothesis, a prediction error response is elicited. This process can be estimated in electroencephalography (EEG) by the mismatch negativity and P3a event related potentials (ERPs) in adults or the mismatch response in infants. It remains unclear is how this system responds to unexpected absence of a sound created by silences. In this thesis, we compared ERPs in adults (Chapter 2) elicited by infrequent sound omissions — i.e. unexpected silences or omission deviants — in various sequences of tones to those elicited by regularly occurring omissions — i.e., expected silences or predictable omissions. We found that unexpected silences elicited both the MMN and P3a, although the magnitude of these components was considerably smaller than we would expected from previous research with omission deviants and auditory deviants. We also found that infants (Chapter 3) exhibited a neural response to omission deviants similar to the mismatch response. Unexpectedly this was not influenced by the global predictability of the omission deviants, which typically attenuates the ERPs to a deviant when it is globally predictable. Adults also showed a lack of difference between globally predictable and globally unpredictable omission deviants (Chapter 4). Furthermore, in adults, we did not find the typical deviance detection ERP responses. Overall, we found evidence of robust neural responses to omission deviants in both adults and infants, but the context in which the omission deviants can change the ERP components elicited. This dissertation is the first to investigate the direct effect of global predictability on the neural responses to omission deviants, as well as 6-month-old infants’ response to omission deviants. / Thesis / Candidate in Philosophy / Often when we expect to hear a sound, instead we “hear” silence or an omission of the sound. This thesis investigates how the brain responds to these unexpected omissions in adults and infants. Unexpected silences elicit a response similar to what we would find after an unexpected change to a sound. This seems to be true for adults and infants as young as 6-months old. Typically, predictable sound changes elicit smaller brain responses. Unlike unexpected sound changes, if the silence occurs predictably in the sequence (e.g., occurs after every 4 tones in a sequence) compared to randomly, or unpredictably, no difference is found. This lack of difference seems to be present in infants and adults. These findings further our understanding of how the brain response to unexpected omissions may not follow the same pattern as the response to unexpected changes of a sound.

In this thesis, robust transmission of 3D models represented by static or time consistent animated meshes is studied from the aspects of scalable coding, multiple description coding (MDC) and error resilient coding. First, three methods for MDC of static meshes are proposed which are based on multiple description scalar quantization, partitioning wavelet trees and optimal protection of scalable bitstream by forward error correction (FEC) respectively. For each method, optimizations and tools to decrease complexity are presented. The FEC based MDC method is also extended as a method for packet loss resilient transmission followed by in-depth analysis of performance comparison with state of the art techniques, which pointed significant improvement. Next, three methods for MDC of animated meshes are proposed which are based on layer duplication and partitioning of the set of vertices of a scalable coded animated mesh by spatial or temporal subsampling where each set is encoded separately to generate independently decodable bitstreams. The proposed MDC methods can achieve varying redundancy allocations by including a number of encoded spatial or temporal layers from the other description. The algorithms are evaluated with redundancy-rate-distortion curves and per-frame reconstruction analysis. Then for layered predictive compression of animated meshes, three novel prediction structures are proposed and integrated into a state of the art layered predictive coder. The proposed structures are based on weighted spatial/temporal prediction and angular relations of triangles between current and previous frames. The experimental results show that compared to state of the art scalable predictive coder, up to 30% bitrate reductions can be achieved with the combination of proposed prediction schemes depending on the content and quantization level. Finally, optimal quality scalability support is proposed for the state of the art scalable predictive animated mesh coding structure, which only supports resolution scalability. Two methods based on arranging the bitplane order with respect to encoding or decoding order are proposed together with a novel trellis based optimization framework. Possible simplifications are provided to achieve tradeoff between compression performance and complexity. Experimental results show that the optimization framework achieves quality scalability with significantly better compression performance than state of the art without optimization.

D’après les théories inférentielles de la perception, notre cerveau tire parti des régularités statistiques présentes dans l’environnement pour générer des prédictions qui façonnent nos contenus perceptifs. Le travail réalisé pendant cette thèse inclut 3 études principales, dans le but de caractériser les déterminants neuronaux des fausses perceptions et la nature des prédictions neuronales.Les erreurs perceptives pourraient résulter d’une tentative de notre système visuel d’expliquer des entrées sensorielles imprécises par une hypothèse erronée. Dans une 1ère étude en Imagerie par Résonance Magnétique fonctionnelle (IRMf), nous montrons que les fausses détections sont associées à un état imprécis et biaisé des circuits sensoriels avant la perception.La répétition d’une image génère une activité neurale diminuée (‘Repetition Suppression’) et des temps de réponse plus courts (‘amorçage’). Ces phénomènes pourraient résulter de mécanismes prédictifs, sous une prédiction implicite de répétition. Dans une 2nde étude IRMf, nous montrons que cette prédiction ne peut pas être modulée par l’expérience, ce qui suggère une implémentation locale. Dans une série d’études comportementales, nous montrons que l’amorçage est modulé par les prédictions, suggérant un mécanisme prédictif. Notre 2nde étude IRMf montre aussi qu’une région de moyen niveau sensible aux visages code des prédictions liées à l’identité, ce qui nous informe sur la nature des prédictions visuelles.Nos résultats montrent que notre perception est façonnée par l’interaction de nos entrées sensorielles avec l’état des circuits neuronaux avant stimulation, qu’il s’agisse de l’activité spontanée ou des stimuli précédents. / According to theoretical frameworks casting perception as inference, our brain can learn the statistical regularities present in the sensory world, and use this prior information to generate predictions, which in turn shape our perceptual contents. The work conducted in this PhD includes three main studies aimed at characterizing the neural determinants of misperceptions, as well as the nature of neural predictions. Perceptual errors may arise from an attempt of our visual system to 'explain' impreciseinputs with an erroneous hypothesis. In a first functional Magnetic Resonance Imaging (fMRI) study, we show that during a detection task, hallucinations are associated with animprecise and biased state of sensory circuits preceding sensation. Stimulus repetition is associated with decreased neural responses, known as Repetition Suppression, and shorter response times, known as priming. These phenomena may reflectpredictive mechanisms under an implicit prior over repetition. In a second fMRI study, we show that this putative prior cannot be changed by experience, suggesting a local, possibly hard-wired neural implementation. In a series of behavioral experiments, we show thatpriming is modulated by predictions, supporting a predictive account of this phenomenon. Our second fMRI study also shows that a mid-level face-sensitive region codes for exemplarspecific predictions, which sheds light on the nature of the predictions encoded along thevisual hierarchy. Altogether, our results speak to the dependence of perception on prior brain states. Both spontaneous activity in sensory circuits and previous stimulation interact with sensory inputsto shape our perceptual contents.