Thalamocortical processing in vision

Abstract

Visual information reaches the cerebral cortex through a major thalamocortical pathway that connects the Lateral Geniculate Nucleus (LGN) of the thalamus with the primary visual area of the cortex (area V1). In humans, ~3.4 million afferents from the LGN are distributed within a V1 surface of ~2,400 mm2, an afferent number that is reduced by half in the macaque and by more than two orders of magnitude in the mouse. Thalamocortical afferents are sorted in visual cortex based on the spatial position of their receptive fields to form a map of visual space. The visual resolution within this map is strongly correlated with total number of thalamic afferents that V1 receives and the area available to sort them. The ~20,000 afferents of the mouse are only sorted by spatial position because they have to cover a large visual field (~300 degrees) within just 4 mm2 of V1 area. By contrast, the ~500,000 afferents of the cat are also sorted by eye input and light/dark polarity because they cover a smaller visual field (~200 degrees) within a much larger V1 area (~ 400 mm2), a sorting principle that is likely to apply also to macaques and humans. The increased precision of thalamic sorting allows building multiple copies of the V1 visual map for left/right eyes and light/dark polarities, which become interlaced to keep neurons representing the same visual point close together. In turn, this interlaced arrangement makes cortical neurons with different preferences for stimulus orientation to rotate around single cortical points forming a pinwheel pattern that allows more efficient processing of objects and visual textures.

Retino-Cortical Mapping Ratio Predicts Columnar and Salt-and-Pepper Organization in Mammalian Visual Cortex

“…Previous studies showed that orientation tuning in V1 is well predicted by the local arrangement of ON and OFF thalamic inputs (Jin et al, 2011;Lien and Scanziani, 2013), implying that functional circuits in V1 might initially be structured by thalamic afferents. In a subsequent study, Mazade and Alonso (2017) also suggested that observed variation of thalamo-cortical projection could play an important role in development of the cortical functions and in maximizing visual acuity. In addition, considering that neurons in the lateral geniculate nucleus (LGN) relay the receptive field of only one to three retinal ganglion cells (RGCs) in most cases (cat: Usrey et al [1999], monkey: Schein and de Monasterio [1987], mouse: Litvina and Chen [2017]), the structure of the LGN afferents reflects that of the retinal feedforward afferents.…”

Origins of Functional Organization in the Visual Cortex

“…It is also not clear how the geometrical realities imposed by different retinal designs might affect the arrangement of ON and OFF RGCs (see below for retinal designs). Mazade and Alonso (2017) propose that the spacing between thalamic axon patches with overlapping receptive fields needs to be greater than 2 axon patches to allow afferents to cluster within different cortical domains. As shown in Figure 4A, as the number of LGN neurons increases, the size of V1 increases in a non-linear fashion, as shown by the exponential plot (y = 770x 1.23 , R 2 = 0.98) (Stevens, 2001), i.e.…”

“…The dotted vertical line shows the possible threshold between species with pinwheels and those without. Reproduced fromMazade and Alonso (2017) with permission from the copyright holder.…”

Sparse long-range connections in visual cortex for cost-efficient small-world networks

“…When the model cortex size was large (gray squirrel 35,38 , 82 mm 2 ), introduction of LRCs increased both classification accuracy and small-worldness (Fig 4c, rightmost). In contrast, when the cortex size was small (mouse 35,39 , 3 mm 2 ), the performance and small-worldness did not increase further regardless of how many LRCs were added (Fig 4c, leftmost). In general, we found a strong correlation between the increment of performance (ΔPerf) and small-worldness (ΔSW) in relation to the variation of V1 size (Fig 4d).…”

“…All the other parameters (including input cell density, sparsity of feedforward connections, and V1 cell density) were assumed to be consistent across species. The anatomical data of averaged V1 size used in this study were from mouse 35,39 (3 mm 2 ), rat 35,54 (7 mm 2 ), tree shrew 50 (73 mm 2 ), and gray squirrel 35,38 (82 mm 2 ). Based on the tree shrew model (784 hidden units) we defined earlier, we set 900 units for gray squirrel, 25 units for the mouse and 64 units for the rat models.…”