Parameter-free binarization and skeletonization of fiber networks from confocal image stacks

Predictive coding and stochastic resonance as fundamental principles of auditory phantom perception

Mechanistic insight is achieved only when experiments are employed to test formal or computational models. Furthermore, in analogy to lesion studies, phantom perception may serve as a vehicle to understand the fundamental processing principles underlying healthy auditory perception. With a special focus on tinnitus-as the prime example of auditory phantom perception-we review recent work at the intersection of artificial intelligence, psychology and neuroscience. In particular, we discuss why everyone with tinnitus suffers from (at least hidden) hearing loss, but not everyone with hearing loss suffers from tinnitus. We argue that intrinsic neural noise is generated and amplified along the auditory pathway as a compensatory mechanism to restore normal hearing based on adaptive stochastic resonance. The neural noise increase can then be misinterpreted as auditory input and perceived as tinnitus. This mechanism can be formalized in the Bayesian brain framework, where the percept (posterior) assimilates a prior prediction (brain's expectations) and likelihood (bottom-up neural signal). A higher mean and lower variance (i.e. enhanced precision) of the likelihood shifts the posterior, evincing a misinterpretation of sensory evidence, which may be further confounded by plastic changes in the brain that underwrite prior predictions. Hence, two fundamental processing principles provide the most explanatory power for the emergence of auditory phantom perceptions: predictive coding as a top-down and adaptive stochastic resonance as a complementary bottom-up mechanism. We conclude that both principles also play a crucial role in healthy auditory perception. Finally, in the context of neuroscience-inspired artificial intelligence, both processing principles may serve to improve contemporary machine learning techniques.

Fiber alignment in 3D collagen networks as a biophysical marker for cell contractility

Cells cultured in 3D fibrous biopolymer matrices exert traction forces on their environment that induce deformations and remodeling of the fiber network. By measuring these deformations, the traction forces can be reconstructed if the mechanical properties of the matrix and the force-free matrix configuration are known. These requirements limit the applicability of traction force reconstruction in practice. In this study, we test whether force-induced matrix remodeling can instead be used as a proxy for cellular traction forces. We measure the traction forces of hepatic stellate cells and different glioblastoma cell lines and quantify matrix remodeling by measuring the fiber orientation and fiber density around these cells. In agreement with simulated fiber networks, we demonstrate that changes in local fiber orientation and density are directly related to cell forces. By resolving Rho-kinase (ROCK) inhibitor-induced changes of traction forces, fiber alignment, and fiber density in hepatic stellate cells, we show that the method is suitable for drug screening assays. We conclude that differences in local fiber orientation and density, which are easily measurable, can be used as a qualitative proxy for changes in traction forces. The method is available as an open-source Python package with a graphical user interface.

Neural network based formation of cognitive maps of semantic spaces and the putative emergence of abstract concepts

How do we make sense of the input from our sensory organs, and put the perceived information into context of our past experiences? The hippocampal-entorhinal complex plays a major role in the organization of memory and thought. The formation of and navigation in cognitive maps of arbitrary mental spaces via place and grid cells can serve as a representation of memories and experiences and their relations to each other. The multi-scale successor representation is proposed to be the mathematical principle underlying place and grid cell computations. Here, we present a neural network, which learns a cognitive map of a semantic space based on 32 different animal species encoded as feature vectors. The neural network successfully learns the similarities between different animal species, and constructs a cognitive map of 'animal space' based on the principle of successor representations with an accuracy of around 30% which is near to the theoretical maximum regarding the fact that all animal species have more than one possible successor, i.e. nearest neighbor in feature space. Furthermore, a hierarchical structure, i.e. different scales of cognitive maps, can be modeled based on multi-scale successor representations. We find that, in fine-grained cognitive maps, the animal vectors are evenly distributed in feature space. In contrast, in coarse-grained maps, animal vectors are highly clustered according to their biological class, i.e. amphibians, mammals and insects. This could be a putative mechanism enabling the emergence of new, abstract semantic concepts. Finally, even completely new or incomplete input can be represented by interpolation of the representations from the cognitive map with remarkable high accuracy of up to 95%. We conclude that the successor representation can serve as a weighted pointer to past memories and experiences, and may therefore be a crucial building block to include prior knowledge, and to derive context knowledge from novel input. Thus, our model provides a new tool to complement contemporary deep learning approaches on the road towards artificial general intelligence.

A systematic review and comparison of automated tools for quantification of fibrous networks

Fibrous networks are essential structural components of biological and engineered materials. Accordingly, many approaches have been developed to quantify their structural properties, which define their material properties. However, a comprehensive overview and comparison of methods is lacking. Therefore, we systematically searched for automated tools quantifying network characteristics in confocal, stimulated emission depletion (STED) or scanning electron microscopy (SEM) images and compared these tools by applying them to fibrin, a prototypical fibrous network in thrombi. Structural properties of fibrin such as fiber diameter and alignment are clinically relevant, since they influence the risk of thrombosis. Based on a systematic comparison of the automated tools with each other, manual measurements, and simulated networks, we provide guidance to choose appropriate tools for fibrous network quantification depending on imaging modality and structural parameter. These tools are often able to reliably measure relative changes in network characteristics, but absolute numbers should be interpreted with care. STATEMENT OF SIGNIFICANCE: Structural properties of fibrous networks define material properties of many biological and engineered materials. Many methods exist to automatically quantify structural properties, but an overview and comparison is lacking. In this work, we systematically searched for all publicly available automated analysis tools that can quantify structural properties of fibrous networks. Next, we compared them by applying them to microscopy images of fibrin networks. We also benchmarked the automated tools against manual measurements or synthetic images. As a result, we give advice on which automated analysis tools to use for specific structural properties. We anticipate that researchers from a large variety of fields, ranging from thrombosis and hemostasis to cancer research, and materials science, can benefit from our work.

Classification at the accuracy limit: facing the problem of data ambiguity

Data classification, the process of analyzing data and organizing it into categories or clusters, is a fundamental computing task of natural and artificial information processing systems. Both supervised classification and unsupervised clustering work best when the input vectors are distributed over the data space in a highly non-uniform way. These tasks become however challenging in weakly structured data sets, where a significant fraction of data points is located in between the regions of high point density. We derive the theoretical limit for classification accuracy that arises from this overlap of data categories. By using a surrogate data generation model with adjustable statistical properties, we show that sufficiently powerful classifiers based on completely different principles, such as perceptrons and Bayesian models, all perform at this universal accuracy limit under ideal training conditions. Remarkably, the accuracy limit is not affected by certain non-linear transformations of the data, even if these transformations are non-reversible and drastically reduce the information content of the input data. We further compare the data embeddings that emerge by supervised and unsupervised training, using the MNIST data set and human EEG recordings during sleep. We find for MNIST that categories are significantly separated not only after supervised training with back-propagation, but also after unsupervised dimensionality reduction. A qualitatively similar cluster enhancement by unsupervised compression is observed for the EEG sleep data, but with a very small overall degree of cluster separation. We conclude that the handwritten letters in MNIST can be considered as 'natural kinds', whereas EEG sleep recordings are a relatively weakly structured data set, so that unsupervised clustering will not necessarily re-cover the human-defined sleep stages.

Fiber finding algorithm using stepwise tracing to identify biopolymer fibers in noisy 3D images

We present a novel fiber finding algorithm (FFA) that will permit researchers to detect and return traces of individual biopolymers. Determining the biophysical properties and structural cues of biopolymers can permit researchers to assess the progression and severity of disease. Confocal microscopy images are a useful method for observing biopolymer structures in three dimensions, but their utility for identifying individual biopolymers is impaired by noise inherent in the acquisition process, including convolution from the point spread function (PSF). The new, iterative FFA we present here 1) measures a microscope's PSF and uses it as a metric for identifying fibers against the background; 2) traces each fiber within a cone angle; and 3) blots out the identified trace before identifying another fiber. Blotting out the identified traces in each iteration allows the FFA to detect and return traces of single fibers accurately and efficiently-even within fiber bundles. We used the FFA to trace unlabeled collagen type I fibers-a biopolymer used to mimic the extracellular matrix in in vitro cancer assays-imaged by confocal reflectance microscopy in three dimensions, enabling quantification of fiber contour length, persistence length, and three-dimensional (3D) mesh size. Based on 3D confocal reflectance microscopy images and the PSF, we traced and measured the fibers to confirm that colder gelation temperatures increased fiber contour length, persistence length, and 3D mesh size-thereby demonstrating the FFA's use in quantifying biopolymers' structural and physical cues from noisy microscope images.

A computational model for understanding the micro-mechanics of collagen fiber network in the tunica adventitia

Abdominal aortic aneurysm is a prevalent cardiovascular disease with high mortality rates. The mechanical response of the arterial wall relies on the organizational and structural behavior of its microstructural components, and thus, a detailed understanding of the microscopic mechanical response of the arterial wall layers at loads ranging up to rupture is necessary to improve diagnostic techniques and possibly treatments. Following the common notion that adventitia is the ultimate barrier at loads close to rupture, in the present study, a finite element model of adventitial collagen network was developed to study the mechanical state at the fiber level under uniaxial loading. Image stacks of the rabbit carotid adventitial tissue at rest and under uniaxial tension obtained using multi-photon microscopy were used in this study, as well as the force-displacement curves obtained from previously published experiments. Morphological parameters like fiber orientation distribution, waviness, and volume fraction were extracted for one sample from the confocal image stacks. An inverse random sampling approach combined with a random walk algorithm was employed to reconstruct the collagen network for numerical simulation. The model was then verified using experimental stress-stretch curves. The model shows the remarkable capacity of collagen fibers to uncrimp and reorient in the loading direction. These results further show that at high stretches, collagen network behaves in a highly non-affine manner, which was quantified for each sample. A comprehensive parameter study to understand the relationship between structural parameters and their influence on mechanical behavior is presented. Through this study, the model was used to conclude important structure-function relationships that control the mechanical response. Our results also show that at loads close to rupture, the probability of failure occurring at the fiber level is up to 2%. Uncertainties in usually employed rupture risk indicators and the stochastic nature of the event of rupture combined with limited knowledge on the microscopic determinants motivate the development of such an analysis. Moreover, this study will advance the study of coupling microscopic mechanisms to rupture of the artery as a whole.

Image-based Characterization of 3D Collagen Networks and the Effect of Embedded Cells

Collagen microstructure is closely related to the mechanical properties of tissues and affects cell migration through the extracellular matrix. To study these structures, three-dimensional (3D) in vitro collagen-based gels are often used, attempting to mimic the natural environment of cells. Some key parameters of the microstructure of these gels are fiber orientation, fiber length, or pore size, which define the mechanical properties of the network and therefore condition cell behavior. In the present study, an automated tool to reconstruct 3D collagen networks is used to extract the aforementioned parameters of gels of different collagen concentration and determine how their microstructure is affected by the presence of cells. Two different experiments are presented to test the functionality of the method: first, collagen gels are embedded within a microfluidic device and collagen fibers are imaged by using confocal fluorescence microscopy; second, collagen gels are directly polymerized in a cell culture dish and collagen fibers are imaged by confocal reflection microscopy. Finally, we investigate and compare the collagen microstructure far from and in the vicinities of MDA-MB 23 cells, finding that cell activity during migration was able to strongly modify the orientation of the collagen fibers and the porosity-related values.

Characterization of the role of collagen network structure and composition in cancer cell migration

The geometry of 3D collagen networks is a key factor that influences the behavior of live cells within extra-cellular matrices. This paper presents a method for automatic quantification of the 3D collagen network geometry with fiber resolution in confocal reflection microscopy images. The proposed method is based on a smoothing filter and binarization of the collagen network followed by a fiber reconstruction algorithm. The method is validated on 3D collagen gels with various collagen and Matrigel concentrations. The results reveal that Matrigel affects the collagen network geometry by decreasing the network pore size while preserving the fiber length and fiber persistence length. The influence of network composition and geometry, especially pore size, is preliminarily analyzed by quantifying the migration patterns of lung cancer cells within microfluidic devices filled with three different hydrogel types. The experiments reveal that Matrigel, while decreasing pore size, stimulates cell migration. Further studies on this relationship could be instrumental for the study of cancer metastasis and other biological processes involving cell migration.

SOAX: a software for quantification of 3D biopolymer networks

Filamentous biopolymer networks in cells and tissues are routinely imaged by confocal microscopy. Image analysis methods enable quantitative study of the properties of these curvilinear networks. However, software tools to quantify the geometry and topology of these often dense 3D networks and to localize network junctions are scarce. To fill this gap, we developed a new software tool called “SOAX”, which can accurately extract the centerlines of 3D biopolymer networks and identify network junctions using Stretching Open Active Contours (SOACs). It provides an open-source, user-friendly platform for network centerline extraction, 2D/3D visualization, manual editing and quantitative analysis. We propose a method to quantify the performance of SOAX, which helps determine the optimal extraction parameter values. We quantify several different types of biopolymer networks to demonstrate SOAX's potential to help answer key questions in cell biology and biophysics from a quantitative viewpoint.

Estimating the 3D pore size distribution of biopolymer networks from directionally biased data

The pore size of biopolymer networks governs their mechanical properties and strongly impacts the behavior of embedded cells. Confocal reflection microscopy and second harmonic generation microscopy are widely used to image biopolymer networks; however, both techniques fail to resolve vertically oriented fibers. Here, we describe how such directionally biased data can be used to estimate the network pore size. We first determine the distribution of distances from random points in the fluid phase to the nearest fiber. This distribution follows a Rayleigh distribution, regardless of isotropy and data bias, and is fully described by a single parameter--the characteristic pore size of the network. The bias of the pore size estimate due to the missing fibers can be corrected by multiplication with the square root of the visible network fraction. We experimentally verify the validity of this approach by comparing our estimates with data obtained using confocal fluorescence microscopy, which represents the full structure of the network. As an important application, we investigate the pore size dependence of collagen and fibrin networks on protein concentration. We find that the pore size decreases with the square root of the concentration, consistent with a total fiber length that scales linearly with concentration.

Image-based three-dimensional analysis to characterize the texture of porous scaffolds

The aim of the present study is to characterize the microstructure of composite scaffolds for bone tissue regeneration containing different ratios of chitosan/gelatin blend and bioactive glasses. Starting from realistic 3D models of the scaffolds reconstructed from micro-CT images, the level of heterogeneity of scaffold architecture is evaluated performing a lacunarity analysis. The results demonstrate that the presence of the bioactive glass component affects not only macroscopic features such as porosity, but mainly scaffold microarchitecture giving rise to structural heterogeneity, which could have an impact on the local cell-scaffold interaction and scaffold performances. The adopted approach allows to investigate the scale-dependent pore distribution within the scaffold and the related structural heterogeneity features, providing a comprehensive characterization of the scaffold texture.

Collective cell traction force analysis on aligned smooth muscle cell sheet between three-dimensional microwalls

During the past two decades, novel biomaterial scaffold for cell attachment and culture has been developed for applications in tissue engineering, biosensing and regeneration medicine. Tissue engineering of blood vessels remains a challenge owing to the complex three-layer histology involved. In order to engineer functional blood vessels, it is essential to recapitulate the characteristics of vascular smooth muscle cells (SMCs) inside the tunica media, which is known to be critical for vasoconstriction and vasodilation of the circulatory system. Until now, there has been a lack of understanding on the mechanotransduction of the SMC layer during the transformation from viable synthetic to quiescent contractile phenotypes. In this study, microfabricated arrays of discontinuous microwalls coated with fluorescence microbeads were developed to probe the mechanotransduction of the SMC layer. First, the system was exploited for stimulating the formation of a highly aligned orientation of SMCs in native tunica medium. Second, atomic force microscopy in combination with regression analysis was applied to measure the elastic modulus of a polyacrylamide gel layer coated on the discontinuous microwall arrays. Third, the conventional traction force assay for single cell measurement was extended for applications in three-dimensional cell aggregates. Then, the biophysical effects of discontinuous microwalls on the mechanotransduction of the SMC layer undergoing cell alignment were probed. Generally, the cooperative multiple cell-cell and cell-microwall interactions were accessed quantitatively by the newly developed assay with the aid of finite-element modelling. The results show that the traction forces of highly aligned cells lying in the middle region between two opposing microwalls were significantly lower than those lying adjacent to the microwalls. Moreover, the spatial distributions of Von Mises stress during the cell alignment process were dependent on the collective cell layer orientation. Immunostaining of the SMC sheet further demonstrated that the collective mechanotransduction induced by three-dimensional topographic cues was correlated with the reduction of actin and vinculin expression. In addition, the online two-dimensional LC-MS/MS analysis verified the modulation of focal adhesion formation under the influence of microwalls through the regulation in the expression of three key cytoskeletal proteins.

3D actin network centerline extraction with multiple active contours

Fluorescence microscopy is frequently used to study two and three dimensional network structures formed by cytoskeletal polymer fibers such as actin filaments and actin cables. While these cytoskeletal structures are often dilute enough to allow imaging of individual filaments or bundles of them, quantitative analysis of these images is challenging. To facilitate quantitative, reproducible and objective analysis of the image data, we propose a semi-automated method to extract actin networks and retrieve their topology in 3D. Our method uses multiple Stretching Open Active Contours (SOACs) that are automatically initialized at image intensity ridges and then evolve along the centerlines of filaments in the network. SOACs can merge, stop at junctions, and reconfigure with others to allow smooth crossing at junctions of filaments. The proposed approach is generally applicable to images of curvilinear networks with low SNR. We demonstrate its potential by extracting the centerlines of synthetic meshwork images, actin networks in 2D Total Internal Reflection Fluorescence Microscopy images, and 3D actin cable meshworks of live fission yeast cells imaged by spinning disk confocal microscopy. Quantitative evaluation of the method using synthetic images shows that for images with SNR above 5.0, the average vertex error measured by the distance between our result and ground truth is 1 voxel, and the average Hausdorff distance is below 10 voxels.