[Cognitive neuro-imaging : phylogenesis and ontogenesis].

Brain imaging methods can visualize the anatomy and activity of children's brains, from the first year of life. These structures and activities are subsequently "recycled" by acquisitions such as reading and arithmetic. Evoked potential mapping of babies aged only a few months, with 64 or 128 electrodes, has a temporal precision of the millisecond order. Magneto-encephalography (MEG) is a costly method. Safe at the field strengths currently used (1.5 or 3 T), MRI requires anti-ferromagnetic precautions, noise protection (tunnel, helmet), and the presence of a parent in the child's field of view. Despite these efforts, movement artifacts remain more of a problem than with adult brain imaging. Brain organization in very young children includes cortical folds and their evolution with time in the premature brain; asymmetry and lateralization of language; the superior temporal sulcus; and anatomical asymmetries combined with early functional organization. The response to spoken language (two-three months), activation of the temporal region and left inferior frontal region (Broca)--as in adults, the asymmetry favoring the left hemisphere during temporal activation (planum). The temporal lobe is hierarchically organized: as in adults, the fastest responses are observed in the primary auditory cortex, decreasing gradually as one progresses to the planum temporal. Syllabes discrimination, regardless of the speaker, and a fusion of visual and auditory information with respect to speech. The newborn brain specializes in the vowels and consonants of the mother tongue, and early visual recognition, especially of faces. Evoked potentials to study the separation of a dorsal track and a classical ventral track in adults. Ventral = identification of objects regardless of their position and orientation; dorsal visual pathway = calculation of the position of objects, action planning, but also counting the number of objects. Evoked potentials at 3 months = brain = already capable of observing changes in the nature and number of a set of objects. We have observed at this early age a clear division between ventral (left temporal) processing of identity and dorsal (right parietal) processing of numbers of objects. The infant brain is already organized during the first months of life. My postulate is that these early cerebral biases provide a framework which constrains cultural learning. Cultural inventions such as reading or symbolic calculation invade the cortical circuits that have evolved in a different context but that can be partly recycled for new human-specific uses. Each cultural object occupies a particular brain niche, a circuit that is already organized but that has sufficient plasticity to be retrained.