bioRxiv ScienceSearch

Biology subjects

Schmidt, A. M.

Publications and source records attributed to Schmidt, A. M..

2 recordsLinked to original sources

Colour Ambiguity In Real Scenes And The Role Of Perceptual Organisation

In this study we show a reproduction of the dress-ambiguity phenomenon in a real scene and we report quantitative measurements of the corresponding colour perceptions. The original, real dress, known from #thedress-illusion, was illuminated by combined short- and longwavelength broadband lights from two slide projectors. Test subjects viewing the dress reported to perceive the dress fabric and lace colours as blue & black, white & gold or light blue & brown; their corresponding perceptual matches were distributed along the blue/yellow cardinal axis, and exhibited a variability comparable to the ambiguity of the dress photograph. It is particularly noteworthy that the colour ambiguity emerged despite explicit knowledge of the observers about the direction of the light source. Manipulating the background of the real dress (change in chromaticity and luminance, or masking) revealed significant differences between the perceptual groups regarding lightness and colour of the dress. Our findings suggest that observer specific differences in the perceptual organisation of the visual scene are responsible for the colour ambiguity observed for the real dress; in particular, we conclude that colour computations of white & gold viewers focused onto the local region of the dress, whereas the colour processes of blue & black and light-blue & brown viewers were strongly influenced by contextual computations including the background. Our segmentation hypothesis extends existing explanations for the dress ambiguity and proposes image based (in the case of the real scene) and high level (in the case of the photograph) neural processes which control the spatial reach of contextual colour computations. The relation between the ambiguity in our real scene and the dress photograph is discussed.

neuroscience

Unbiased metabolic flux inference through combined thermodynamic and 13C flux analysis

ABSTRACTQuantification of cellular metabolic fluxes, for instance with 13C-metabolic flux analysis, is highly important for applied and fundamental metabolic research. A current challenge in 13C-flux analysis is that the available experimental data are usually insufficient to resolve metabolic fluxes in large metabolic networks without making assumptions on flux directions and reversibility. To infer metabolic fluxes in a more unbiased manner, we devised an approach that does not require such assumptions. The developed three-step approach integrates thermodynamics, metabolome, physiological data, and 13C labelling data, and involves a novel method to comprehensively sample the complex thermodynamically-constrained metabolic flux space. Applying our approach to budding yeast with its compartmentalised metabolism and parallel pathways, we could resolve metabolic fluxes in an unbiased manner, we obtained an uncertainty estimate for each flux, and we found novel flux patterns that until now had remained unknown, likely due to assumptions made in previous 13C flux analysis studies. We expect that our approach will be an important step forward to determine metabolic fluxes with improved accuracy in microorganisms and possibly also in more complex organisms.View Full Text

systems biology