Neural representations of social and non-social uncertainty in human decision making

The social landscape is filled with an intricate web of species-specific desired objects and course of actions. Humans are highly social animals and, as they navigate this landscape, they need to produce adapted decision-making behaviour. Traditionally social and non-social neural mechanisms affecting choice have been investigated using different approaches. Recently, in an effort to unite these findings, two main theories have been proposed to explain how the brain might encode social and non-social motivational decision-making: the extended common currency and the social valuation specific schema (Ruff & Fehr 2014). One way to test these theories is to directly compare neural activity related to social and non-social decision outcomes within the same experimental setting. Here we address this issue by focusing on the neural substrates of social and non-social forms of uncertainty. Using functional magnetic resonance imaging (fMRI) we directly compared the neural representations of reward and risk prediction and errors (RePE and RiPE) in social and non- social situations using gambling games. We used a trust betting game to vary uncertainty along a social dimension (trustworthiness), and a card game (Preuschoff et al. 2006) to vary uncertainty along a non-social dimension (pure risk). The trust game was designed to maintain the same structure of the card game. In a first study, we exposed a divide between subcortical and cortical regions when comparing the way these regions process social and non-social forms of uncertainty during outcome anticipation. Activity in subcortical regions reflected social and non-social RePE, while activity in cortical regions correlated with social RePE and non-social RiPE. The second study focused on outcome delivery and integrated the concept of RiPE in non-social settings with that of fairness and monetary utility maximisation in social settings. In particular these results corroborate recent models of anterior insula function (Singer et al. 2009; Seth 2013), and expose a possible neural mechanism that weights fairness and uncertainty but not monetary utility. The third study focused on functionally defined regions of the early visual cortex (V1) showing how activity in these areas, traditionally considered only visual, might reflect motivational prediction errors in addition to known perceptual prediction mechanisms (den Ouden et al 2012). On the whole, while our results do not support unilaterally one or the other theory modeling the underlying neural dynamics of social and non-social forms of decision making, they provide a working framework where both general mechanisms might coexist.

Deploying and processing neural representations of signals

Neural representations (NR) have emerged in the last few years as a powerful tool to represent signals from several domains, such as images, 3D shapes, or audio. Indeed, deep neural networks have been shown capable of approximating continuous functions that describe a given signal with theoretical infinite resolution. This finding allows obtaining representations whose memory footprint is fixed and decoupled from the resolution at which the underlying signal can be sampled, something that is not possible with traditional discrete representations, e.g., grids of pixels for images or voxels for 3D shapes. During the last two years, many techniques have been proposed to improve the capability of NR to approximate high-frequency details and to make the optimization procedures required to obtain NR less demanding both in terms of time and data requirements, motivating many researchers to deploy NR as the main form of data representation for complex pipelines. Following this line of research, we first show that NR can approximate precisely Unsigned Distance Functions, providing an effective way to represent garments that feature open 3D surfaces and unknown topology. Then, we present a pipeline to obtain in a few minutes a compact Neural Twin® for a given object, by exploiting the recent advances in modeling neural radiance fields. Furthermore, we move a step in the direction of adopting NR as a standalone representation, by considering the possibility of performing downstream tasks by processing directly the NR weights. We first show that deep neural networks can be compressed into compact latent codes. Then, we show how this technique can be exploited to perform deep learning on implicit neural representations (INR) of 3D shapes, by only looking at the weights of the networks.

Location-independent and location-linked representations of sound objects.

For the recognition of sounds to benefit perception and action, their neural representations should also encode their current spatial position and their changes in position over time. The dual-stream model of auditory processing postulates separate (albeit interacting) processing streams for sound meaning and for sound location. Using a repetition priming paradigm in conjunction with distributed source modeling of auditory evoked potentials, we determined how individual sound objects are represented within these streams. Changes in perceived location were induced by interaural intensity differences, and sound location was either held constant or shifted across initial and repeated presentations (from one hemispace to the other in the main experiment or between locations within the right hemispace in a follow-up experiment). Location-linked representations were characterized by differences in priming effects between pairs presented to the same vs. different simulated lateralizations. These effects were significant at 20-39 ms post-stimulus onset within a cluster on the posterior part of the left superior and middle temporal gyri; and at 143-162 ms within a cluster on the left inferior and middle frontal gyri. Location-independent representations were characterized by a difference between initial and repeated presentations, independently of whether or not their simulated lateralization was held constant across repetitions. This effect was significant at 42-63 ms within three clusters on the right temporo-frontal region; and at 165-215 ms in a large cluster on the left temporo-parietal convexity. Our results reveal two varieties of representations of sound objects within the ventral/What stream: one location-independent, as initially postulated in the dual-stream model, and the other location-linked.

Shared neural circuits for mentalizing about the self and others

Although many examples exist for shared neural representations of self and other, it is unknown how such shared representations interact with the rest of the brain. Furthermore, do high-level inference-based shared mentalizing representations interact with lower level embodied/simulation-based shared representations? We used functional neuroimaging (fMRI) and a functional connectivity approach to assess these questions during high-level inference-based mentalizing. Shared mentalizing representations in ventromedial prefrontal cortex, posterior cingulate/precuneus, and temporo-parietal junction (TPJ) all exhibited identical functional connectivity patterns during mentalizing of both self and other. Connectivity patterns were distributed across low-level embodied neural systems such as the frontal operculum/ventral premotor cortex, the anterior insula, the primary sensorimotor cortex, and the presupplementary motor area. These results demonstrate that identical neural circuits are implementing processes involved in mentalizing of both self and other and that the nature of such processes may be the integration of low-level embodied processes within higher level inference-based mentalizing.

Plasticity in the neural coding of auditory space in the mammalian brain

Sound localization relies on the neural processing of monaural and binaural spatial cues that arise from the way sounds interact with the head and external ears. Neurophysiological studies of animals raised with abnormal sensory inputs show that the map of auditory space in the superior colliculus is shaped during development by both auditory and visual experience. An example of this plasticity is provided by monaural occlusion during infancy, which leads to compensatory changes in auditory spatial tuning that tend to preserve the alignment between the neural representations of visual and auditory space. Adaptive changes also take place in sound localization behavior, as demonstrated by the fact that ferrets raised and tested with one ear plugged learn to localize as accurately as control animals. In both cases, these adjustments may involve greater use of monaural spectral cues provided by the other ear. Although plasticity in the auditory space map seems to be restricted to development, adult ferrets show some recovery of sound localization behavior after long-term monaural occlusion. The capacity for behavioral adaptation is, however, task dependent, because auditory spatial acuity and binaural unmasking (a measure of the spatial contribution to the “cocktail party effect”) are permanently impaired by chronically plugging one ear, both in infancy but especially in adulthood. Experience-induced plasticity allows the neural circuitry underlying sound localization to be customized to individual characteristics, such as the size and shape of the head and ears, and to compensate for natural conductive hearing losses, including those associated with middle ear disease in infancy.

Sound objects in time, space and action

SOUND OBJECTS IN TIME, SPACE AND ACTIONThe term "sound object" describes an auditory experience that is associated with an acoustic event produced by a sound source. At cortical level, sound objects are represented by temporo-spatial activity patterns within distributed neural networks. This investigation concerns temporal, spatial and action aspects as assessed in normal subjects using electrical imaging or measurement of motor activity induced by transcranial magnetic stimulation (TMS).Hearing the same sound again has been shown to facilitate behavioral responses (repetition priming) and to modulate neural activity (repetition suppression). In natural settings the same source is often heard again and again, with variations in spectro-temporal and spatial characteristics. I have investigated how such repeats influence response times in a living vs. non-living categorization task and the associated spatio-temporal patterns of brain activity in humans. Dynamic analysis of distributed source estimations revealed differential sound object representations within the auditory cortex as a function of the temporal history of exposure to these objects. Often heard sounds are coded by a modulation in a bilateral network. Recently heard sounds, independently of the number of previous exposures, are coded by a modulation of a left-sided network.With sound objects which carry spatial information, I have investigated how spatial aspects of the repeats influence neural representations. Dynamics analyses of distributed source estimations revealed an ultra rapid discrimination of sound objects which are characterized by spatial cues. This discrimination involved two temporo-spatially distinct cortical representations, one associated with position-independent and the other with position-linked representations within the auditory ventral/what stream.Action-related sounds were shown to increase the excitability of motoneurons within the primary motor cortex, possibly via an input from the mirror neuron system. The role of motor representations remains unclear. I have investigated repetition priming-induced plasticity of the motor representations of action sounds with the measurement of motor activity induced by TMS pulses applied on the hand motor cortex. TMS delivered to the hand area within the primary motor cortex yielded larger magnetic evoked potentials (MEPs) while the subject was listening to sounds associated with manual than non- manual actions. Repetition suppression was observed at motoneuron level, since during a repeated exposure to the same manual action sound the MEPs were smaller. I discuss these results in terms of specialized neural network involved in sound processing, which is characterized by repetition-induced plasticity.Thus, neural networks which underlie sound object representations are characterized by modulations which keep track of the temporal and spatial history of the sound and, in case of action related sounds, also of the way in which the sound is produced.LES OBJETS SONORES AU TRAVERS DU TEMPS, DE L'ESPACE ET DES ACTIONSLe terme "objet sonore" décrit une expérience auditive associée avec un événement acoustique produit par une source sonore. Au niveau cortical, les objets sonores sont représentés par des patterns d'activités dans des réseaux neuronaux distribués. Ce travail traite les aspects temporels, spatiaux et liés aux actions, évalués à l'aide de l'imagerie électrique ou par des mesures de l'activité motrice induite par stimulation magnétique trans-crânienne (SMT) chez des sujets sains. Entendre le même son de façon répétitive facilite la réponse comportementale (amorçage de répétition) et module l'activité neuronale (suppression liée à la répétition). Dans un cadre naturel, la même source est souvent entendue plusieurs fois, avec des variations spectro-temporelles et de ses caractéristiques spatiales. J'ai étudié la façon dont ces répétitions influencent le temps de réponse lors d'une tâche de catégorisation vivant vs. non-vivant, et les patterns d'activité cérébrale qui lui sont associés. Des analyses dynamiques d'estimations de sources ont révélé des représentations différenciées des objets sonores au niveau du cortex auditif en fonction de l'historique d'exposition à ces objets. Les sons souvent entendus sont codés par des modulations d'un réseau bilatéral. Les sons récemment entendus sont codé par des modulations d'un réseau du côté gauche, indépendamment du nombre d'expositions. Avec des objets sonores véhiculant de l'information spatiale, j'ai étudié la façon dont les aspects spatiaux des sons répétés influencent les représentations neuronales. Des analyses dynamiques d'estimations de sources ont révélé une discrimination ultra rapide des objets sonores caractérisés par des indices spatiaux. Cette discrimination implique deux représentations corticales temporellement et spatialement distinctes, l'une associée à des représentations indépendantes de la position et l'autre à des représentations liées à la position. Ces représentations sont localisées dans la voie auditive ventrale du "quoi".Des sons d'actions augmentent l'excitabilité des motoneurones dans le cortex moteur primaire, possiblement par une afférence du system des neurones miroir. Le rôle des représentations motrices des sons d'actions reste peu clair. J'ai étudié la plasticité des représentations motrices induites par l'amorçage de répétition à l'aide de mesures de potentiels moteurs évoqués (PMEs) induits par des pulsations de SMT sur le cortex moteur de la main. La SMT appliquée sur le cortex moteur primaire de la main produit de plus grands PMEs alors que les sujets écoutent des sons associée à des actions manuelles en comparaison avec des sons d'actions non manuelles. Une suppression liée à la répétition a été observée au niveau des motoneurones, étant donné que lors de l'exposition répétée au son de la même action manuelle les PMEs étaient plus petits. Ces résultats sont discuté en termes de réseaux neuronaux spécialisés impliqués dans le traitement des sons et caractérisés par de la plasticité induite par la répétition. Ainsi, les réseaux neuronaux qui sous-tendent les représentations des objets sonores sont caractérisés par des modulations qui gardent une trace de l'histoire temporelle et spatiale du son ainsi que de la manière dont le son a été produit, en cas de sons d'actions.

Étude multimodale de la consolidation d’habiletés motrices à ‎l’aide de l’IRMf et de l’EEG

La consolidation est le processus qui transforme une nouvelle trace mnésique labile en ‎une autre plus stable et plus solide. Une des tâches utilisées en laboratoire pour ‎l’exploration de la consolidation motrice dans ses dimensions comportementale et ‎cérébrale est la tâche d’apprentissage de séquences motrices. Celle-ci consiste à ‎reproduire une même série de mouvements des doigts, apprise de manière implicite ou ‎explicite, tout en mesurant l’amélioration dans l’exécution. Les études récentes ont ‎montré que, dans le cas de l’apprentissage explicite de cette tâche, la consolidation de la ‎trace mnésique associée à cette nouvelle habileté dépendrait du sommeil, et plus ‎particulièrement des fuseaux en sommeil lent. Et bien que deux types de fuseaux aient ‎été décrits (lents et rapides), le rôle de chacun d’eux dans la consolidation d’une ‎séquence motrice est encore mal exploré. En effet, seule une étude s’est intéressée à ce ‎rôle, montrant alors une implication des fuseaux rapides dans ce processus mnésique ‎suite à une nuit artificiellement altérée. D’autre part, les études utilisant l’imagerie ‎fonctionnelle (IRMf et PET scan) menées par différentes équipes dont la notre, ont ‎montré des changements au niveau de l’activité du système cortico-striatal suite à la ‎consolidation motrice. Cependant, aucune corrélation n’a été faite à ce jour entre ces ‎changements et les caractéristiques des fuseaux du sommeil survenant au cours de la nuit ‎suivant un apprentissage moteur. Les objectifs de cette thèse étaient donc: 1) de ‎déterminer, à travers des enregistrements polysomnographiques et des analyses ‎corrélationnelles, les caractéristiques des deux types de fuseaux (i.e. lents et rapides) ‎associées à la consolidation d’une séquence motrice suite à une nuit de sommeil non ‎altérée, et 2) d’explorer, à travers des analyses corrélationnelles entre les données ‎polysomnographiques et le signal BOLD (« Blood Oxygenated Level Dependent »), ‎acquis à l’aide de l’imagerie par résonance magnétique fonctionnelle (IRMf), ‎l’association entre les fuseaux du sommeil et les activations cérébrales suite à la ‎consolidation de la séquence motrice. Les résultats de notre première étude ont montré ‎une implication des fuseaux rapides, et non des fuseaux lents, dans la consolidation ‎d’une séquence motrice apprise de manière explicite après une nuit de sommeil non ‎altérée, corroborant ainsi les résultats des études antérieures utilisant des nuits de ‎sommeil altérées. En effet, les analyses statistiques ont mis en évidence une ‎augmentation significative de la densité des fuseaux rapides durant la nuit suivant ‎l’apprentissage moteur par comparaison à la nuit contrôle. De plus, cette augmentation ‎corrélait avec les gains spontanés de performance suivant la nuit. Par ailleurs, les ‎résultats de notre seconde étude ont mis en évidence des corrélations significatives entre ‎l’amplitude des fuseaux de la nuit expérimentale d’une part et les gains spontanés de ‎performance ainsi que les changements du signal BOLD au niveau du système cortico-‎striatal d’autre part. Nos résultats suggèrent donc un lien fonctionnel entre les fuseaux ‎du sommeil, les gains de performance ainsi que les changements neuronaux au niveau ‎du système cortico-striatal liés à la consolidation d’une séquence motrice explicite. Par ‎ailleurs, ils supportent l’implication des fuseaux rapides dans ce type de consolidation ; ‎ceux-ci aideraient à l’activation des circuits neuronaux impliqués dans ce processus ‎mnésique et amélioreraient par la même occasion la consolidation motrice liée au ‎sommeil.‎

The acquisition of skilled motor performance: Fast and slow experience-driven changes in primary motor cortex

Behavioral and neurophysiological studies suggest that skill learning can be mediated by discrete, experience-driven changes within specific neural representations subserving the performance of the trained task. We have shown that a few minutes of daily practice on a sequential finger opposition task induced large, incremental performance gains over a few weeks of training. These gains did not generalize to the contralateral hand nor to a matched sequence of identical component movements, suggesting that a lateralized representation of the learned sequence of movements evolved through practice. This interpretation was supported by functional MRI data showing that a more extensive representation of the trained sequence emerged in primary motor cortex after 3 weeks of training. The imaging data, however, also indicated important changes occurring in primary motor cortex during the initial scanning sessions, which we proposed may reflect the setting up of a task-specific motor processing routine. Here we provide behavioral and functional MRI data on experience-dependent changes induced by a limited amount of repetitions within the first imaging session. We show that this limited training experience can be sufficient to trigger performance gains that require time to become evident. We propose that skilled motor performance is acquired in several stages: “fast” learning, an initial, within-session improvement phase, followed by a period of consolidation of several hours duration, and then “slow” learning, consisting of delayed, incremental gains in performance emerging after continued practice. This time course may reflect basic mechanisms of neuronal plasticity in the adult brain that subserve the acquisition and retention of many different skills.

On cortical coding of vocal communication sounds in primates

Understanding how the brain processes vocal communication sounds is one of the most challenging problems in neuroscience. Our understanding of how the cortex accomplishes this unique task should greatly facilitate our understanding of cortical mechanisms in general. Perception of species-specific communication sounds is an important aspect of the auditory behavior of many animal species and is crucial for their social interactions, reproductive success, and survival. The principles of neural representations of these behaviorally important sounds in the cerebral cortex have direct implications for the neural mechanisms underlying human speech perception. Our progress in this area has been relatively slow, compared with our understanding of other auditory functions such as echolocation and sound localization. This article discusses previous and current studies in this field, with emphasis on nonhuman primates, and proposes a conceptual platform to further our exploration of this frontier. It is argued that the prerequisite condition for understanding cortical mechanisms underlying communication sound perception and production is an appropriate animal model. Three issues are central to this work: (i) neural encoding of statistical structure of communication sounds, (ii) the role of behavioral relevance in shaping cortical representations, and (iii) sensory–motor interactions between vocal production and perception systems.

Associative knowledge controls deployment of visual selective attention

According to some models of visual selective attention, objects in a scene activate corresponding neural representations, which compete for perceptual awareness and motor behavior. During a visual search for a target object, top-down control exerted by working memory representations of the target's defining properties resolves competition in favor of the target. These models, however, ignore the existence of associative links among object representations. Here we show that such associations can strongly influence deployment of attention in humans. In the context of visual search, objects associated with the target were both recalled more often and recognized more accurately than unrelated distractors. Notably, both target and associated objects competitively weakened recognition of unrelated distractors and slowed responses to a luminance probe. Moreover, in a speeded search protocol, associated objects rendered search both slower and less accurate. Finally, the first saccades after onset of the stimulus array were more often directed toward associated than control items.

Learning joint representations for order and timing of perceptual-motor sequences: a dynamic neural field approach

Many of our everyday tasks require the control of the serial order and the timing of component actions. Using the dynamic neural field (DNF) framework, we address the learning of representations that support the performance of precisely time action sequences. In continuation of previous modeling work and robotics implementations, we ask specifically the question how feedback about executed actions might be used by the learning system to fine tune a joint memory representation of the ordinal and the temporal structure which has been initially acquired by observation. The perceptual memory is represented by a self-stabilized, multi-bump activity pattern of neurons encoding instances of a sensory event (e.g., color, position or pitch) which guides sequence learning. The strength of the population representation of each event is a function of elapsed time since sequence onset. We propose and test in simulations a simple learning rule that detects a mismatch between the expected and realized timing of events and adapts the activation strengths in order to compensate for the movement time needed to achieve the desired effect. The simulation results show that the effector-specific memory representation can be robustly recalled. We discuss the impact of the fast, activation-based learning that the DNF framework provides for robotics applications.

Learning task specific distributed paragraph representations using a 2-tier convolutional neural network

We introduce a type of 2-tier convolutional neural network model for learning distributed paragraph representations for a special task (e.g. paragraph or short document level sentiment analysis and text topic categorization). We decompose the paragraph semantics into 3 cascaded constitutes: word representation, sentence composition and document composition. Specifically, we learn distributed word representations by a continuous bag-of-words model from a large unstructured text corpus. Then, using these word representations as pre-trained vectors, distributed task specific sentence representations are learned from a sentence level corpus with task-specific labels by the first tier of our model. Using these sentence representations as distributed paragraph representation vectors, distributed paragraph representations are learned from a paragraph-level corpus by the second tier of our model. It is evaluated on DBpedia ontology classification dataset and Amazon review dataset. Empirical results show the effectiveness of our proposed learning model for generating distributed paragraph representations.

Governing coordination: Behavioural principles and neural correlates

The coordination of movement is governed by a coalition of constraints. The expression of these constraints ranges from the concrete—the restricted range of motion offered by the mechanical configuration of our muscles and joints; to the abstract—the difficulty that we experience in combining simple movements into complex rhythms. We seek to illustrate that the various constraints on coordination are complementary and inclusive, and the means by which their expression and interaction are mediated systematically by the integrative action of the central nervous system (CNS). Beyond identifying the general principles at the behavioural level that govern the mutual interplay of constraints, we attempt to demonstrate that these principles have as their foundation specific functional properties of the cortical motor systems. We propose that regions of the brain upstream of the motor cortex may play a significant role in mediating interactions between the functional representations of muscles engaged in sensorimotor coordination tasks. We also argue that activity in these ldquosupramotorrdquo regions may mediate the stabilising role of augmented sensory feedback.

Effect of levodopa on both verbal and motor representations of action in Parkinson's disease: a fMRI study.

Previous studies have demonstrated that non-demented Parkinson's disease (PD) patients have a specific impairment of verb production compared with noun generation. One interpretation of this deficit suggested the influence of striato-frontal dysfunction on action-related verb processing. The aim of our study was to investigate cerebral changes after motor improvement due to dopaminergic medication on the neural circuitry supporting action representation in the brain as mediated by verb generation and motor imagery in PD patients. Functional magnetic resonance imaging on 8 PD patients in "ON" dopaminergic treatment state (DTS) and in "OFF" DTS was used to explore the brain activity during three different tasks: Object Naming (ObjN), Generation of Action Verbs (GenA) in which patients were asked to overtly say an action associated with a picture and mental simulation of action (MSoA) was investigated by asking subjects to mentally simulate an action related to a depicted object. The distribution of brain activities associated with these tasks whatever DTS was very similar to results of previous studies. The results showed that brain activity related to semantics of action is modified by dopaminergic treatment in PD patients. This cerebral reorganisation concerns mainly motor and premotor cortex suggesting an involvement of the putaminal motor loop according to the "motor" theory of verb processing.