Statistical image properties predict aesthetic ratings in abstract paintings created by neural style transfer.

Artificial intelligence has emerged as a powerful computational tool to create artworks. One application is Neural Style Transfer, which allows to transfer the style of one image, such as a painting, onto the content of another image, such as a photograph. In the present study, we ask how Neural Style Transfer affects objective image properties and how beholders perceive the novel (style-transferred) stimuli. In order to focus on the subjective perception of artistic style, we minimized the confounding effect of cognitive processing by eliminating all representational content from the input images. To this aim, we transferred the styles of 25 diverse abstract paintings onto 150 colored random-phase patterns with six different Fourier spectral slopes. This procedure resulted in 150 style-transferred stimuli. We then computed eight statistical image properties (complexity, self-similarity, edge-orientation entropy, variances of neural network features, and color statistics) for each image. In a rating study, we asked participants to evaluate the images along three aesthetic dimensions (Pleasing, Harmonious, and Interesting). Results demonstrate that not only objective image properties, but also subjective aesthetic preferences transferred from the original artworks onto the style-transferred images. The image properties of the style-transferred images explain 50 - 69% of the variance in the ratings. In the multidimensional space of statistical image properties, participants considered style-transferred images to be more Pleasing and Interesting if they were closer to a "sweet spot" where traditional Western paintings (JenAesthetics dataset) are represented. We conclude that NST is a useful tool to create novel artistic stimuli that preserve the image properties of the input style images. In the novel stimuli, we found a strong relationship between statistical image properties and subjective ratings, suggesting a prominent role of perceptual processing in the aesthetic evaluation of abstract images.

Fibronectin is a smart adhesive that both influences and responds to the mechanics of early spinal column development.

An extracellular matrix of Fibronectin adheres the neural tube to the two flanking columns of paraxial mesoderm and is required for normal vertebrate development. Here, we find that the bilaterally symmetric interfaces between the zebrafish neural tube and paraxial mesoderm function as optimally engineered adhesive lap joints with rounded edges, graded Fibronectin 'adhesive' and an arced adhesive spew filet. Fibronectin is a 'smart adhesive' that remodels to the lateral edges of the neural tube-paraxial mesoderm interfaces where shear stress is highest. Fibronectin remodeling is mechanically responsive to contralateral variation morphogenesis, and Fibronectin-mediated inter-tissue adhesion is required for bilaterally symmetric morphogenesis of the paraxial mesoderm. Strikingly, however, perturbation of the Fibronectin matrix rescues the neural tube convergence defect of cadherin 2 mutants. Therefore, Fibronectin-mediated inter-tissue adhesion dynamically coordinates bilaterally symmetric morphogenesis of the vertebrate trunk but predisposes the neural tube to convergence defects that lead to spina bifida.

In embryos, the spinal cord starts out as a flat sheet of cells that curls up to form a closed cylinder called the neural tube. The folding tube is attached to the surrounding tissues through an extracellular matrix of proteins and sugars. Overlapping strands of a protein from the extracellular matrix called Fibronectin connect the neural tube to adjacent tissues, like a kind of biological glue. However, it remained unclear what effect this attachment had on the embryonic development of the spinal cord. Connecting two overlapping objects with glue to form what is known as an 'adhesive lap joint' is common in fields such as woodworking and aeronautical engineering. The glue in these joints comes under shearing stress whenever the two objects it connects try to pull apart. But, thanks to work in engineering, it is possible to predict how different joints will perform under tension. Now, Guillon et al. have deployed these engineering principles to shed light on neural tube development. Using zebrafish embryos and computational models, Guillon et al. investigated what happens when the strength of the adhesive lap joints in the developing spine changes. This revealed that Fibronectin works like a smart adhesive: rather than staying in one place like a conventional glue, it moves around. As the neural tube closes, cells remodel the Fibronectin, concentrating it on the areas under the highest stress. This seemed to both help and hinder neural tube development. On the one hand, by anchoring the tube equally to the left and right sides of the embryo, the Fibronectin glue helped the spine to develop symmetrically. On the other hand, the strength of the adhesive lap joints made it harder for the neural tube to curl up and close. If the neural tube fails to close properly, it can lead to birth defects like spina bifida. One of the best-known causes of these birth defects in humans is a lack of a vitamin known as folic acid. Cell culture experiments suggest that this might have something to do with the mechanics of the cells during development. It may be that faulty neural tubes could close more easily if they were able to unglue themselves from the surrounding tissues. Further use of engineering principles could shed more light on this idea in the future.

The application of inelastic neutron scattering to investigate the interaction of methyl propanoate with silica.

A modern industrial route for the manufacture of methyl methacrylate involves the reaction of methyl propanoate and formaldehyde over a silica-supported Cs catalyst. Although the process has been successfully commercialised, little is known about the surface interactions responsible for the forward chemistry. This work concentrates upon the interaction of methyl propanoate over a representative silica. A combination of infrared spectroscopy, inelastic neutron scattering, DFT calculations, X-ray diffraction and temperature-programmed desorption is used to deduce how the ester interacts with the silica surface.