Usually different materials have their specific natural textures with bumps and dents, such as fabric, fur, stone and leather and on the bases of this texture they can be distinguished with one another. In Japanese concept of “shitsukan” this perception is known as “NATURAL TEXTURE PERCEPTION”. In this blog, I am going to discuss about about the neural mechanisms underlying material’s texture perception, or shitsukan perception with special focus on “MATERIAL TEXTURE PERCEPTION”.
Usually different materials have their specific natural textures with bumps and dents, such as fabric, fur, stone and leather and on the bases of this texture they can be distinguished with one another. This specif surface textures also influence the way light is reflected along the surface and this influence is different for different materials. For example, in case of a furry object, thin fiber-like structures are arranged in a specific flow pattern, and the anisotropic light reflections from each fiber, as well as inter-reflection between fibers, are repeated. Likewise in case of fabrics, light is reflected from individual fibers depending on their optical properties at a microscopic scale and their geometries. At the same time, light reflection changes in accordance with the pattern of the woven structure on a mesoscopic scale, which results in a texture specific to the fabric. In materials like wood and some kinds of stone, such as granite, which consists of multiple components with different reflective properties, repetitive changes in color and luminescence generates material-specific textures. In materials like sand or sugar, which are accumulations of large numbers of small grains, even if the reflective properties of each grain are simple, the repetition of shadings due to the 3D structure of each grain produces specific textures. In all these cases, textures are characterized by incompleteness of the regularity with random fluctuations that are distinct from artificial, abstract textures on flat surface, that have been traditionally used as visual stimuli. Another important characteristic of natural textures resides in their regularity and the randomness.In addition the form of image statistics plays an important role in the processing of natural textures. In fact, images that give the same impression as the original texture have been successfully constructed by matching specific kinds of image statistics. Of particular interest are algorithms that involve extraction of local orientations and spatial frequencies using Gabor-like filters early during processing, which is similar to image processing within the visual cortex.
Neural processing of natural texture in humans
The higher areas along the ventral pathway in humans reside in the ventral visual cortex. Evidence from neuroimaging studies indicates that, within this region, information about natural textures is processed to a large extent in parallel with shape information. It is found that attending to natural texture (judging differences in texture) activates cortical regions around the collateral sulcus (CoS) in the medial part of the ventral visual cortex. Attending to the object shape (judging differences in shape), by contrast, activates more lateral regions known to be a shape–selective area (lateral occipital, LO). It is evidenced by several studies that information about natural texture is processed in the medial part of the ventral visual cortex around the CoS largely independently of object shape information, which is processed more laterally; red symbols in Figure below. The CoS in humans is a sulcus running in the posterior-anterior direction through the ventral visual cortex (white lines in Figure below). Around the CoS, the parahippocampal gyrus (PHG) is situated anteromedially, the lingual gyrus (LG) posteromedially, and the fusiform gyrus(FG) laterally. The studies proved that the texture/material-related activities are not localized within a single region. Instead, they are widely distributed along the CoS, from the vicinity of its posterior end to more anterior parts and extend into adjacent gyri (PHG, LG, and FG) in the medial–lateral direction.
The functional differences between these texture/material-related regions along the CoS, particularly those between the posterior and anterior parts, are not fully understood yet. A series of studies showed that regions selective to natural texture in the medial part overlap with a scene-selective sub-region called the parahippocampal place area (PPA), and this region is sensitive to object ensembles (scenes composed of lots of objects) (yellow circles in above Figure). This joint sensitivity to natural texture and scenes such as object ensembles would be related to the fact that both natural texture and scene perception use image statistics information. The image statistics for scenes are called “scene gist,”. That part of the PPA is thus likely to be responsible for computation of image statistics features important for scene and texture perception. In macaques, it has been reported that many neurons within scene-selective sub-regions in the IT cortex, potential homologous to the human PPA, exhibit response selectivity to natural textures. Although what image features these neurons respond to remains unresolved, it has been suggested some image components such as high spatial frequency or long straight lines play an important role. On the other hand, the FG region lateral to the CoS and medial to the shape-selective LO often exhibits response selectivity for both natural texture and shape. This texture-sensitive region in the FG has some overlap with the object-selective subregion (pFs, see also section “Neural responses to gloss: humans”) and the face-selective subregion. A region that is jointly selective for natural texture, shape, and color has also been reported in the FG (green square in above Figure). Although it remains uncertain, this region may host integrated representations of objects by combining information about natural texture, color, and shape, each of which is separately analyzed around it.
In my next blogs I will elaborate more about the third important modality “RECOGNITION OF MATERIAL CATEGORY AND ITS PROPERTIES”.
- Komatsu H, Goda N. Neural mechanisms of material perception: Quest on Shitsukan. Neuroscience. 2018 Nov 10;392:329-47.