On the relationship between maps and domains in inferotemporal cortex

Abstract

How does the brain encode information about the environment? Decades of research have led to the pervasive notion that the object-processing pathway in primate cortex consists of multiple areas that are each specialized to process different object categories (such as faces, bodies, hands, non-face objects and scenes). The anatomical consistency and modularity of these regions have been interpreted as evidence that these regions are innately specialized. Here, we propose that ventral-stream modules do not represent clusters of circuits that each evolved to process some specific object category particularly important for survival, but instead reflect the effects of experience on a domain-general architecture that evolved to be able to adapt, within a lifetime, to its particular environment. Furthermore, we propose that the mechanisms underlying the development of domains are both evolutionarily old and universal across cortex. Topographic maps are fundamental, governing the development of specializations across systems, providing a framework for brain organization.

Fig. 1. Development of face selectivity in macaque and human infants.

Lack of face versus non-face object selectivity in macaque and human infants. a | Before approximately 200 days old, macaques do not show face > non-face-object selective regions, and after this age, face selectivity appears and is stable, as measured by functional MRI. b | Cerebral blood volume signal responses reveal that, before face and non-face-object domains become detectable, monkey inferotemporal cortex is responsive to visual stimuli, but not selective to image category. c | Regions that are face selective in adult human ventral occipital temporal cortex are not selective for faces over non-face objects in 4–6-month-old human infants, as reflected by percentage differences in blood oxygen level-dependent responses Some selectivity for scenes versus faces was observed in both these studies, but could reflect differences in visual-field stimulation (for example, centre versus periphery) by the stimuli used. FDR, false discovery rate; ROI, region of interest. Parts a and b adapted from REF., CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Part c adapted from REF., CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/).

Fig. 2. Intrinsic and experience-dependent organization in macaque visual cortex.

a | Lack of face domains (left) but normal hand domains (right) in face-deprived monkeys (bottom row) as compared with controls (top row). Images show contrasts as determined with functional MRI. b | At birth, most of the cortex is made up of maps of the sensory periphery. This is a composite image illustrating the coverage of visual, somatosensory–motor and auditory maps. Maps of eccentricity in visual space cover occipital, temporal, posterior parietal and frontal eye fields. Maps of the body (face, hands and feet) cover areas within and around the central sulcus. Alternating representations of high (yellow, with white outline) and low (cyan, with black outline) tonotopic frequencies cover parts of the superior temporal gyrus. Part a adapted from REF., Springer Nature Limited. Part b adapted with permission from REF., PNAS, and from REF., Springer Nature Limited, using data from REFS^,.

Fig. 3. Topographic receptive-field tuning.

a | Curvature correlates with eccentricity throughout the macaque visual system. Maps of responses to peripheral minus central-visual-field stimuli (left), straight minus curvy stimuli (centre) and non-face objects minus faces (right), with examples of each stimulus type above each map. A, C, and P indicate anterior, central and posterior subdivisions of inferotemporal cortex, respectively. b | Curvature also correlates with eccentricity in the human visual system The dashed white line represents the border between central and peripheral visual field in early visual cortex. The cyan–yellow scale correlates with curvilinear values of visual stimuli, such that the areas in red–yellow process curvy features and those in blue–cyan process rectilinear features. The locations of the fusiform face area (FFA; green outline), occipital face area (OFA; blue outline), occipital curvature preference patch (OCP; black outline) and fusiform curvature preference patch (FCP; white outline) are within areas that preferentially respond to curved features. By contrast, the location of the parahippocampal place area (PPA) is encompassed by the region responding preferentially to rectilinear features. The probabilistic locations of face domains (the OFA and FFA) and the scene domain (the PPA) are superimposed on a human eccentricity map (right). c | Gradient of shape selectivity in macaque inferotemporal cortex that could reflect developmental origins in eccentricity-based low-level shape selectivity gradients. Colours show the correspondence between clustering in image shape space (top) and anatomical space (bottom). A, anterior; D, dorsal; P, posterior; RSC, retrosplenial cortex; STS, superior temporal sulcus; V, ventral; V1, primary visual cortex. Part a adapted from REF., Springer Nature Limited. Part b adapted with permission from REF., Elsevier (left) and generated using data from REFS^,. Part c adapted from REF., Springer Nature Limited.

Fig. 4. Congruence between sensory maps.

a | Global congruency of sensory map orientation in mice (left), macaques (middle) and humans (right). b | Congruence of visual and somatosensory maps in parietal cortex Neurons responding to stimulation of central visual space also respond to touch on central parts of the face, whereas neurons responding to peripheral visual space respond to touch on more peripheral body parts. c | Connectivity between early visual cortex and inferotemporal cortex is predominantly along isoeccentricities. Neuronal tracer injection sites in central-visual-field parts of posterior inferotemporal cortex (dark red) are selectively connected to central-visual-field parts of early visual areas and. higher visual areas (lighter red). Injection sites in peripheral-visual-field parts of intermediate visual areas (dark green) are selectively connected to peripheral-visual-field parts of both lower and higher visual areas (light green). S1, primary somatosensory cortex; V1, primary visual cortex. Part b adapted with permission from REF., The American Physiological Society. Part c adapted with permission from REF., OUP.