Funded Grants

Exploring the organization of the infant brain with functional imaging

Cognitive capacities, such as language, mathematics, music, etc. are highly developed in humans as compared to animals. Numerous studies have found precursors of these capacities in infants. For example, infants are able to discriminate sentences in different languages, distinguish sets of objects based on their numerosity or recognize known faces. These abilities are not very different from those of other animals. Monkeys are also able to discriminate two human languages, two quantities of items, or respond to particular faces. In a few years, however, children surpass these animals. To explain the development of the cognitive capacities of our species, two hypotheses are classically opposed: one postulates a particular brain organization that dedicates specific brain regions to these functions while the other postulates powerful general learning mechanisms, greater than those of other species thanks to a our larger brain and a rich cultural environment.

To reconcile these opposing positions, a perhaps more plausible hypothesis considers that exposure to the human environment may take advantage of preexisting bias in human brain functional architecture. It is thus not necessary to postulate that the human genome codes directly for the human complex cognitive capacities but rather that it codes for initial biases in a cascade of biological events whose result is human cognition. Such an hypothesis is more compatible with recent descriptions of the massive overlap between the human genome and that of other mammals.

The goal of this project is thus to explore the initial organization of the human brain with non invasive brain imaging techniques and to compare the responses of infants' and adults' brains to ecological stimuli that are significant for humans. Because of the recent application of brain imaging techniques in infants, this type of projects needs to pursue in parallel methodological and scientific questions. At a methodological level, the goal is to develop fMRI as a tool to explore infant's brain to complete high-density ERP. A first step will be to rely on demonstrated behavioral infants' capacities to refine fMRI experimental paradigms while investigating the cerebral bases of these capacities. A second step will be to study aspects of cognition that are not easily accessible to infants' behavioral paradigms, that are limited to novelty or familiarity overt responses.

To study the functional organization of the brain, I propose to systematically exploit habituation or priming phenomenon, that is the fact that neuronal response to a stimulus decreases with repetition. This phenomenon has been described at the single cell level but also at a macroscopic level in ERP recordings and fMRI. By manipulating what counts as a repetition, we can infer the code of the representation. For example in neonates, ERP amplitude decreases with syllable repetition but recovers after a change of phoneme. This is observed when the same physical syllable is repeated but also when different speakers produce the syllables. This demonstrates that a representation of syllables, speaker independent, is computed by infants as well as by adults.

Two directions will be pursued in this project:

1) Cerebral bases of infants' linguistic abilities:

To investigate the extent of functional specification for speech in infants and the similarities and differences with adults, two series of experiments are planned, one exploring the sentence level and the other the phoneme level.

A) Although sentence comprehension is certainly very different in three-month-old
infants and adults, the same left temporo-parietal network is involved in both populations. To clarify the different components of this network and their function, the same sentence (one every 14s) will be repeated several times, either identically or with the prosody flattened (intonation is removed) or hummed (phonology is removed). It is expected that fMRI activation will decrease in a network that is repeatedly involved in the same process. Identifying the exact condition under which a given brain area habituates will allow to specify what parameter each area is sensitive to.

B) Based on the differences in topography of mismatch responses elicited by a change of stimuli in series of repeated auditory stimuli, I have suggested that the auditory system is organized very early on in several functional networks, coding the different properties of a sound. The exact topographical localization of these networks and their specificity to linguistic stimuli will be studied using the same A A A X paradigm with fast-event fMRI and high-density ERP.

2) What are the spontaneous categories computed by infants ?

In this second part of the project, I will use the habituation method to probe the anatomical bases and degree of abstraction of mental representations that are thought to be available in infants: faces, numbers, objects, etc.. For example, consider infants sensitivity to faces. Is it related to the particular visual characteristics of human faces, or to take an extreme opposite position, do faces already evoke an abstract representation of conspecifics whose specifications are not limited to face, but include a representation of the body, the voice, etc.? If it is the case, habituation to pictures of body parts should prime responses to faces and induce a decrease of activation in the networks computing conspecific recognition. By manipulating the distance between the context (pictures or sounds) and the target faces, it should be possible to characterize the degree of abstraction of infant's early knowledge of conspecifics. If this methodology proves productive, as suggested by our preliminary ERP results, the same experimental process would be applied to explore infants' knowledge in other domains, such as the relations between shapes, tools and actions, or number and space in ERP and fMRI.

Through these experiments, I will ascertain whether several methodological paradigms in fMRI (blocks, slow and fast-event-related designs, priming method) successful in adults, are useful in infants while exploring how the early brain organization constrains infants' representations of the external world. Bridging brain, mind and behavior requires not only the study of the mature adult brain but also an understanding of the initial stages of brain organization. Progress in functional neuro-imaging provides tools to explore this question and to constrain models of cognitive development.