Information encoded in volumes and areas of dendritic spines is nearly maximal across mammalian brains

Long-term information associated with neuronal memory resides in dendritic spines. However, spines can have a limited size due to metabolic and neuroanatomical constraints, which should effectively limit the amount of encoded information in excitatory synapses. This study investigates how much information can be stored in the sizes of dendritic spines, and whether is it optimal in any sense? It is shown here, using empirical data for several mammalian brains across different regions and physiological conditions, that dendritic spines nearly maximize entropy contained in their volumes and surface areas for a given mean size. This result is essentially independent of the type of a fitting distribution to size data, as both short- and heavy-tailed distributions yield similar nearly 100 % information efficiency in the majority of cases, although heavy-tailed distributions slightly better fit the data. On average, the highest information is contained in spine volume, and the lowest in spine length or spine head diameter. Depending on a species and brain region, a typical spine can encode between 6.1 and 10.8 bits of information in its volume, and 3.1-8.1 bits in its surface area. Our results suggest a universality of entropy maximization in spine volumes and areas, which can be a new principle of memory storing in synapses.

Tinnitus-like “hallucinations” elicited by sensory deprivation in an entropy maximization recurrent neural network

Sensory deprivation has long been known to cause hallucinations or “phantom” sensations, the most common of which is tinnitus induced by hearing loss, affecting 10–20% of the population. An observable hearing loss, causing auditory sensory deprivation over a band of frequencies, is present in over 90% of people with tinnitus. Existing plasticity-based computational models for tinnitus are usually driven by homeostatic mechanisms, modeled to fit phenomenological findings. Here, we use an objective-driven learning algorithm to model an early auditory processing neuronal network, e.g., in the dorsal cochlear nucleus. The learning algorithm maximizes the network’s output entropy by learning the feed-forward and recurrent interactions in the model. We show that the connectivity patterns and responses learned by the model display several hallmarks of early auditory neuronal networks. We further demonstrate that attenuation of peripheral inputs drives the recurrent network towards its critical point and transition into a tinnitus-like state. In this state, the network activity resembles responses to genuine inputs even in the absence of external stimulation, namely, it “hallucinates” auditory responses. These findings demonstrate how objective-driven plasticity mechanisms that normally act to optimize the network’s input representation can also elicit pathologies such as tinnitus as a result of sensory deprivation.

Numerical Investigation on the Effect of the Oxymethylene Ether-3 (OME3) Blending Ratio in Premixed Sooting Ethylene Flames

Synthetic fuels, especially oxygenated fuels, which can be used as blending components, make it possible to modify the emission properties of conventional fossil fuels. Among oxygenated fuels, one promising candidate is oxymethylene ether-3 (OME3). In this work, the sooting propensity of ethylene (C2H4) blended with OME3 is numerically investigated on a series of laminar burner-stabilized premixed flames with increasing amounts of OME3, from pure ethylene to pure OME3. The numerical analysis is performed using the Conditional Quadrature Method of Moments combined with a detailed physico-chemical soot model. Two different equivalence ratios corresponding to a lightly and a highly sooting flame condition have been investigated. The study examines how different blending ratios of the two fuels affect soot particle formation and a correlation between OME3 blending ratio and corresponding soot reduction is established. The soot precursor species in the gas-phase are analyzed along with the soot volume fraction of small nanoparticles and large aggregates. Furthermore, the influence of the OME3 blending on the particle size distribution is studied applying the entropy maximization concept. The effect of increasing amounts of OME3 is found to be different for soot nanoparticles and larger aggregates. While OME3 blending significantly reduces the amount of larger aggregates, only large amounts of OME3, close to pure OME3, lead to a considerable suppression of nanoparticles formed throughout the flame. A linear correlation is identified between the OME3 content in the fuel and the reduction in the soot volume fraction of larger aggregates, while smaller blending ratios may lead to an increased number of nanoparticles for some positions in the flame for the richer flame condition.

Framboid Self-Assembly and Self-Organization

The formation of framboids involves two distinct processes. First, pyrite microcrystals aggregate into spherical groups through surface free energy minimization. The self-assembly of framboid microcrystals to form framboids is consistent with estimations based on the classical Derjaguin-Landau-Verwey-Overbeek (DVLO) theory, which balances the attraction between particles due to the van der Waals forces against the interparticle electrostatic repulsive force. Second, the microcrystals rearrange themselves into ordered domains through entropy maximization. Icosahedral symmetry tends to minimize short-range attractive interactions and maximize entropy. The physical processes which facilitate this rearrangement are Brownian motion and surface interactions. Curved framboid interface enforce deviation from the cubic close packed structure. In the absence of a curved surface, weakly interacting colloidal particles preferentially self-assemble into a cubic close packed structure, and this is observed in irregular, non-framboidal aggregates of pyrite micro- and nanocrystals.

Framboids

Framboids may be the most astonishing and abundant natural features you have never heard of. These microscopic spherules of golden pyrite consist of thousands of even smaller microcrystals, often arranged in stunning geometric arrays. There are probably 1030 on Earth, and they are forming at a rate of 1020 every second. This means that there are a billion times more framboids than sand grains on Earth, and a million times more framboids than stars in the observable universe. They are all around us: they can be found in rocks of all ages and in present-day sediments, soils, and natural waters. The sulfur in the pyrite is mainly produced by bacteria, and many framboids contain organic matter. They are formed through burst nucleation of supersaturated solutions of iron and sulfide, followed by limited crystal growth in diffusion-dominated stagnant sediments. The framboids self-assemble as surface free energy is minimized and the microcrystals are attracted to each other by surface forces. Self-organization occurs through entropy maximization, and the microcrystals rotate into their final positions through Brownian motion. The final shape of the framboids is often actually polygonal or partially facetted rather than spherical, as icosahedral microcrystal packing develops. Their average diameter is around 6 microns and the average microcrystal size is about 0.1 microns. There is no significant change in these dimensions with time: the framboid is an exceptionally stable structure, and the oldest may be 2.9 billion years old. This means that they provide samples of the chemistry of ancient environments.

The Crystallography of Pyrite Framboids

Single crystal X-ray diffraction analyses of even the most perfectly organized framboids show ring patterns indicative of randomly oriented particles. Therefore, framboids are not mesocrystals or extreme skeletal varieties of single crystals. Electron backscatter diffraction shows that the microcrystals within a framboid are not crystallographically aligned. Around half of the microcrystals in organized framboids have crystallographic orientations rotated 90º. The results of single crystal XRD and framboid EBSD studies clearly show that the microcrystals are self-organized rather than being the result of a crystallographic template or preexisting structural control. The pre-formed framboid microcrystals which are initially randomly organized throughout the framboid volume then, in some cases, begin to wholly or partly self-order. This is effected by rotation of the microcrystals until an ordered array is produced. The consequence of this rotation must be that the microcrystals are initially packed loosely enough for rotation to occur. The processes involved in the rotation could include forces intrinsic to the microcrystals themselves, such as surface forces, or forces imposed from outside the framboid, such as Brownian motion. The fundamental driving force for microcrystal rotation and the development of organized microcrystal arrays in framboids is entropy maximization.

Lung Cancer Diagnosis Based on an ANN Optimized by Improved TEO Algorithm

A quarter of all cancer deaths are due to lung cancer. Studies show that early diagnosis and treatment of this disease are the most effective way to increase patient life expectancy. In this paper, automatic and optimized computer-aided detection is proposed for lung cancer. The method first applies a preprocessing step for normalizing and denoising the input images. Afterward, Kapur entropy maximization is performed along with mathematical morphology to lung area segmentation. Afterward, 19 GLCM features are extracted from the segmented images for the final evaluations. The higher priority images are then selected for decreasing the system complexity. The feature selection is based on a new optimization design, called Improved Thermal Exchange Optimization (ITEO), which is designed to improve the accuracy and convergence abilities. The images are finally classified into healthy or cancerous cases based on an optimized artificial neural network by ITEO. Simulation is compared with some well-known approaches and the results showed the superiority of the suggested method. The results showed that the proposed method with 92.27% accuracy provides the highest value among the compared methods.