Projects

Focal Research Area A: Cerebral Mechanics

A01: In silico modelling of brain malformations

Recent evidence suggests that mechanical forces drive cortical folding during brain development. While analytical, computational, and experimental models have significantly advanced our understanding of the mechanisms underlying the folding process, such models have so far not been used to tackle specific clinical challenges. For example, folding patterns are an important clinical hallmark of cortical development and brain malformations, such as those related to epilepsy. Assisting clinicians in the diagnosis and treatment of brain folding-related neurological disorders, e.g., by reverse-engineering the cellular processes that could have led to the macroscopic malformation, requires the close collaboration between clinicians, who define open clinical questions, and engineers, who develop targeted simulation tools to address them. To tackle some of the open questions concerning malformations associated with epilepsy analyzed in A02, we plan to use mechanical modelling approaches. A01 aims to develop a computational framework to numerically predict the mechanisms underlying abnormal brain development. Based on qualitative and quantitative insights into the interplay between mechanics, cell migration, cell differentiation, and brain malformations obtained through projects A02 through A05, we will establish a multifield theoretical framework to predict brain development under physiological and pathological conditions. For model calibration and validation, we will use the various human data sets generated in project A02 supplemented by mechanical tests under compression, tension, and torsional shear on fresh human brain tissue obtained from neurosurgical procedures. We will incorporate data from patients into the modelling framework to lay the foundation for later use in clinical practice – to advance from benchmark problems to disease-specific predictions and eventually assist the diagnosis and treatment of neurological diseases such as epilepsy.

Project leader: Dr.-Ing. Silvia Budday

Position: 1 doctoral researcher

A02: Quantitative characterisation of brain malformations

Abnormal brain folding is a frequent cause of focal seizures in patients with drug-resistant epilepsy, i.e., due to Focal Cortical Dysplasia (FCD), Hemimegalencephaly, Polymicrogyria, or Mild Malformation of Cortical Development with Oligodendroglial Hyperplasia (MOGHE). We and others have identified brain somatic mutations acquired during brain development as underlying genetic cause of these malformations of cortical development (MCD). It remains to be clarified, however, how these molecular-genetic alterations translate into abnormal tissue structure and function. Our working hypothesis postulates that brain mechanics play a pivotal role in the translation of molecular-genetic lesions into abnormal cerebral structure and function. Access to human brain tissue and neuroimaging data of post-mortem body donors and histopathologically and genetically characterised epilepsy surgery brain specimens is a key feature of A02, and will complement the well established animal models and cell cultures available in this CRC. A02 will apply quantitative histological and immunohistochemical measures assisted by enzyme-linked immunosorbent assays, ultrastructural analysis, deep learning algorithms, and spatial transcriptomics to comprehensively define glial and neuronal subtypes and the composition of the extracellular matrix (ECM) in healthy and diseased human brain samples. In addition, we will obtain multimodal high-resolution magnetic resonance imaging data of human brain malformations to identify radiomic, imaging-derived biomarkers that correlate with our anatomo-pathological and molecular findings and visco-elastic features measured in brain tissue from the same patients (together with A01). A key task during the first phase is, therefore, to generate unprecedented neuroanatomical, neuropathological, molecular, neuroimaging and mechanical data of the human brain that are critical for the development of new and comprehensive models of brain malformations in silico (A01), in vitro (A03 and A04), and in vivo (A05).

Project leaders: Prof. Dr. med. Ingmar Blümcke, Prof. Dr. med. Arnd Dörfler, Prof. Dr. med. Friedrich Paulsen

Positions: 2 doctoral researchers

A03: In vitro model for the mechanics of early brain development

A03 aims to investigate the interplay of mechanical factors and planar cell polarity (PCP) signaling in early nervous system development in the African Clawed Frog Xenopus laevis. We plan to develop an organoid model of neural plate tissue by replacing the endogenous supporting tissues with engineered substrates from X03 with tunable mechanical properties, such as stiffness and stress relaxation, and functionalisation with extracellular matrix proteins such as fibronectin. These organoids will allow us to systematically vary the mechanical environment of the developing brain and monitor its morphogenesis and patterning in response to these variations. In addition, we will perturb PCP signaling in developing brain organoids and analyze mechanical properties, cell shape, adhesion and migration, tissue geometry as well as gene expression patterns. From this data, we will develop a model, which describes the influence of mechanics on early brain patterning, morphogenesis and differentiation and integrates mechanics and biochemical signals. In this project we aim to provide in-depth characterisation and experimental manipulation of the mechanical environment in early brain development of Xenopus laevis using an organoid model. Further, we intend to connect tissue mechanics and tissue-scale biochemical signaling, specifically planar cell polarity (PCP) signaling. Within the consortium, this project investigates the earliest stages of brain development, which set the basic patterns and morphological structures for subsequent differentiation. Thereby, it closes the gap between tissue specification and differentiation of the more complex structures of the adult brain and provides relevant data for in silico models, e.g. for A01. In addition, the organoid model developed in this project can be utilised to test for early effects of genes or phenomena investigated by other PLs in the consortium. If successful, this project will provide a versatile model system and novel and relevant insights into the cross-talk of tissue mechanics and a highly conserved and broadly relevant biochemical signaling pathway. Deciphering this cross-talk in the early development of the frog brain will generate important insights on the interaction of physical and biochemical cues and furthermore, may reveal basic and conserved mechanisms in tissue development.

Project leader: Prof. Dr. rer. nat. Alexandra Schambony

Position: 1 doctoral researcher

 

A04: The role of mechanics in orchestrating neural lineage decisions

A04 aims to uncover the role of mechanics in steering human neural stem cell lineage decisions, neuron formation, and neuronal migration in brain organoids, a 3D model for early human brain development derived from induced pluripotent stem cells. Artificial brain tissue engineered to quantitatively tune the physical properties of the environment brain organoids are exposed to will be used to determine the effects of mechanical properties on cellular processes governing brain development. Characterisation of tissue composition will reveal the aggregated effect of the impact on proliferation, differentiation, migration, and the organisation of brain organoid-resident cells. Time-lapse imaging following single neural stem cells and their progeny will be used to temporally resolve the changes induced on a cellular level and allows to reveal the influence of a given physical environment on NSC lineage decisions and alterations in neuronal migration patterns. To uncover the molecular framework relaying changes in the physical environment to alterations in cellular behaviour we will employ single cell RNA-sequencing and functionally assess associated key nodes in developmental neurogenesis.

Project leaders: Prof. Dr. rer. nat. Marisa Karow/Dr. Sven Falk

Position: 1 doctoral researcher

A05: In vivo model for the mechanics of brain development

A05 will use Xenopus laevis to study the mechanobiology of brain development in vivo. First, we will combine time-lapse fluorescence imaging of growing axons with a time-lapse in vivo atomic force microscopy (tiv-AFM) approach, which we propose to develop here to map tissue viscoelasticity at cellular resolution. We will then induce disease-associated mutations in candidate genes identified in A02, such as SLC35A2 or genes associated with the mTOR pathway, and investigate how such genetic perturbations impact on tissue mechanics and cell motility in the developing brain. To test whether phenotypes associated with the genetic perturbations are due to cell-intrinsic processes or because of altered signaling by the environment of the cells, we will complement our in vivo experiments by in vitro studies of cell motility. Healthy and perturbed cells will be acutely isolated from embryonic tissue and cultured on custom-built compliant substrates of varying viscoelasticity to assess the effect of substrate mechanics on neuronal growth. Finally, we will exploit genetic, pharmacological, and physical approaches to manipulate brain mechanics, and test if restoring tissue viscoelasticity to physiological values is sufficient to rescue phenotypes that would normally be caused by genetic mutations associated with brain malformations. Within the timeframe of this 4-years project, we aim to develop a method enabling highly resolved in vivo measurements of brain viscoelasticity.  We will apply it to healthy developing frog brains and brains with mutations known to cause brain malformations in humans, compare the results, and test if rescuing brain mechanics rescues phenotypes associated with brain malformations. The findings of this study will generate important data justifying in vivo approaches. Thus, data acquired here will be used in projects A01 and X01, while the method developed here will be of importance to projects A03, A04, and C02-C05. Brain mechanics could, in addition to well-established genetic and chemical signaling, play a major role in brain malformations. Illuminating the role of genes identified in human brain malformations in the development of the frog brain may reveal conserved signalling pathways. If re-establishing healthy brain mechanics in mutated animals in vivo prevents pathological phenotypes, this project may lead to a much better understanding of brain malformations and ultimately to the development of new treatment strategies.

Project leader: Prof. Dr. Kristian Franze

Position: 1 doctoral researcher

 

Focal Research Area B: Spinal Mechanics

B01: In silico modelling of spinal cord regeneration

B01 aims to establish a continuum-based computational framework to predict the regeneration of spinal cord tissue after injury or disease. The computational model will capture the temporal and spatial evolution of growth, remodelling and healing processes, as experimentally observed in B02-B05. We will specifically focus on mechanics-driven processes that are involved in the regeneration of the spinal cord after traumatic injury and in multiple sclerosis. We will capture the evolving connectivity of cells in the central nervous system by continuous order parameters driven by mechanics and biochemical factors. To calibrate the constitutive models, we will exploit mechanical tests on human and animal spinal cord tissue performed in B01-B05. Modelling in B01 will help correlating the comprehensive set of ex vivo and in vivo mechanical data, data from various species, and multiple measurement techniques within EBM. B01 will in particular enter into a close feedback loop with B03, which provides data on in vivo tissue mechanics based on Brillouin microscopy (BM) measurements and ex vivo tissue mechanics based on atomic force microscopy (AFM) measurements and correlated structural and compositional information. B01 will in turn provide information about mechanical determinants identified through modelling and simulation. B01 will furthermore provide testable hypotheses for targeted experiments in B03 and the associated results will be directly fed back into our computational framework.

Project leaders: Prof. Dr.-Ing. habil Paul Steinmann, Dr.-Ing. Silvia Budday

Positions: 2 doctoral researchers

B02: Pre and post metamorphosis spinal cord regeneration in frogs

Frogs may regenerate CNS neurons before but not after metamorphosis. While some differences in the biochemical landscape of injured frog tissue before and after metamorphosis have already been identified, they cannot fully explain the tremendous differences in the regenerative capacity of neurons, suggesting that other signals may contribute to regulating wound healing and neuronal regeneration. B02 will study mechanical, cellular and molecular changes of spinal cord tissue in the African clawed frog Xenopus laevis before and after metamorphosis and determine which combination of parameters are responsible for the loss of neuronal regeneration in post-metamorphotic froglets.  Tissue mechanics will be assessed by atomic force microscopy and Brillouin microscopy, differences in the genetic and chemical state of cells and the extracellular matrix will be measured using single cell RNAseq, HCR, immunohistochemistry, and proteomics. We will perturb tissue mechanics and gene expression of identified candidate genes in the post metamorphosis spinal cord to ultimately facilitate neuronal regeneration in the adult spinal cord, which would normally fail. Within the timeframe of this 4-years project, we aim to characterise the mechanical landscape of the Xenopus spinal cord before and after injury in regenerative and non-regenerative stages.  We will compare the results with published data on non-regenerative mammalian spinal cords and regenerative zebrafish spinal cords and test if manipulating spinal cord mechanics in non-regenerative, postmitotic froglets rescues neuronal regeneration. The findings of this study will generate important data justifying in vivo approaches. Data acquired here will be of great importance to projects B01 and X01, the method applied here will be extensively used in many other B projects. Spinal cord mechanics could, in addition to well-established genetic and chemical signaling, play a major role in regulating regeneration. Identifying cellular and extracellular components critical for tissue stiffness and neuronal regeneration may reveal conserved signaling pathways. If re-establishing regenerative spinal cord mechanics in non-regenerative froglets in vivo facilitates neuronal regeneration, this project may lead to a much better understanding of spinal cord injuries. It is intriguing to speculate if our findings in Xenopus can be translated into human physiology and contribute to novel treatment approaches in clinical practice.

Project leader: Prof. Dr. Kristian Franze

Position: 1 doctoral researcher

B03: The determinants of spinal cord mechanics in homeostasis

B03 aims at mapping the mechanical landscape of the zebrafish spinal cord by systematically quantifying its mechanical tissue properties in vivo, deciphering their determinants, and identifying their relevance during tissue homeostasis. In particular, we will focus on the extracellular matrix and the cellular contribution to mechanical tissue properties. By employing biochemical and optogenetic tools to induce the targeted deposition of ECM components or reduction of cell bodies, for instance, we will alter the composition and structure of the spinal cord in living zebrafish. We will perform Brillouin microscopy measurements on these specimens in vivo, and investigate the same conditions and manipulations ex vivo by employing both Brillouin microscopy and AFM-enabled indentation measurements on freshly sectioned tissue. Thereby, we aim to bridge multiple spatio-temporal regimes as implemented by the various projects in this consortium, and facilitate potential inference of in vivo mechanical properties from ex vivo measurements. We will correlate the results with structural and compositional analyses of the spinal parenchyma based on histological, microscopic and genetic tools using machine learning. Ultimately, our results will not only contribute to a dataset in which a distinct mechanical phenotype of the spinal cord tissue will be predictable from histological data and proteome analysis. This project will also help facilitate an understanding of the mechanical factors that have to be addressed in order to change a pathological outcome after injury or disease, ideally without affecting other biochemical factors.

Project leaders: Prof. Dr. Jochen Guck/Dr. rer. nat. Stephanie Möllmert

Position: 1 doctoral researcher

B04: Spinal cord mechanics in a mouse model of multiple sclerosis

Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system (CNS) with unclear aetiology. The disease typically starts with a relapsing-remitting phenotype (RRMS), which is mainly triggered by inflammation and the invasion of immune cells from the periphery, before most patients enter secondary progressive MS (SPMS) in the course of the disease. The time point for this transition is unpredictable and shows high interpatient variability. While the reasons for the transition also remain to be elucidated, there is the notion that the peripheral inflammatory response becomes secondary over time and is replaced by CNS-intrinsic mechanisms. Considering such CNS-intrinsic mechanisms, glia cells, in particular astrocytes, may play a crucial role by providing mechanical support to the CNS tissue. The involvement of mechanical forces in CNS-intrinsic disease-driving mechanisms in MS has hardly been addressed so far. In close interaction with projects B03 and B05, project B04 aims to understand how local mechanical properties of the CNS contribute to the pathology of MS. We will measure the evolution of mechanical properties of the spinal cord during the course of experimental autoimmune encephalomyelitis (EAE), which is the most common mouse model of MS. We will evaluate whether these mechanical changes are associated with disease onset and severity, and related to microglia activation, astrogliosis and nerve fiber pathology. In addition, we will focus on mechanosensation by astrocytes through the ion channel Piezo1 and assess the effects on spinal cord pathology when astrocytes lose their ability to respond to mechanical cues.

Project leader: Prof. Dr. med. Stefanie Kürten

Position: 1 postdoctoral researcher

B05: In vivo mechanical manipulation of spinal cord regeneration

Based on the hypothesis that local mechanical properties are critical for axonal regrowth after spinal cord injury, B05 will use Brillouin microscopy to non-invasively map the in vivo mechanical properties of the growth-permissive spinal lesion site in zebrafish; a vertebrate species that combines high regenerative capacity for the central nervous system with optical accessibility. Additionally, atomic force microscopy-based nanointendation measurements (AFM) will be employed to determine the tissue mechanics of the spinal lesion site on living tissue sections. These results will be complemented by a panel of transcriptomic, proteomic, in vivo microscopic and histological analyses of injury-induced changes of extracellular matrix composition and structure. Building on this platform, we will use an array of genetic tools to manipulate the extracellular matrix and investigate its role as a potential key determinant of the mechanical properties of the zebrafish spinal lesion site. We will correlate the mechanical read-outs with regenerative success and establish a link between the mechanical and biochemical properties of the lesion site and axon regeneration. Finally, in collaboration with X03, we will capitalise on the insights obtained from our functional experiments to engineer artificial biomaterials with biochemical and mechanical properties similar to the zebrafish spinal lesion site, to confirm their axon growth-stimulating properties and evaluate their potential for regenerative medicine applications.

Project leader: Dr. rer. nat. Daniel Wehner

Position: 1 doctoral researcher

 

Focal Research Area C: Cellular Mechanics

C01: In silico modelling of mechanical cell-matrix interactions

C01 establishes experimentally informed theoretical modelling, algorithmic implementation and computer simulations to uncover the role of mechanical cell-matrix interactions in the context of brain tissue. Neuronal cells exert active stresses on the surrounding extracellular matrix, thereby also remodelling it, which in turn feedbacks on cell polarity and migration. Agent-based and coarse-grained, phenomenological continuum models accounting for cell-matrix interactions, cell migration and active stress generation will capture these processes in the context of neuronal tissue. Via implementation into in silico models, we will predict, in feedback with C03 and C05, how cell-cell and cell-matrix interactions drive the in vitro hippocampal cell-network formation. In cooperation with A04 and C05, we will furthermore study the dynamics of neuronal organoid formation in artificial matrices based on a nonlinear continuum formulation and its finite element implementation. In doing so, we will establish an in silico set of tools enabling to uncover microscopic mechanisms and interactions otherwise not accessible to in vitro and in vivo measurements and predict the response to biological and mechanical perturbations. Finally, the mechanical properties resulting from cell-cell and cell-matrix interactions identified by our approach will serve as a relevant input for the continuum modelling and simulations in A01 and B01. As our long-term vision, we will also extend our modelling approaches to capture processes of development and regeneration in neuronal tissue.

Project leaders: Prof. Dr.-Ing. habil. Paul Steinmann, Prof. Dr. Vasily Zaburdaev

Positions: 2 doctoral researchers

C02: The role of mechanics for neuronal ‘plasticity’

Activity-dependent structural rearrangements of the neuronal network are fundamental for correct wiring in juvenile brain. Upon their establishment, the neuronal network needs to be stabilised to guarantee circuit functional fidelity, which is achieved by the formation of a perineuronal extracellular matrix (ECM) during adolescence. The perineuronal ECM is a mesh-like structure with a backbone composed of hyaluronic acid and chondroitin sulphate proteoglycans, interacting with various secreted and membrane-associated molecules. This phase of brain development coincides with changes in mechanical properties of brain tissue, which may be due to the physical properties of the perineuronal ECM. Thus, ECM-derived changes in the mechanical properties of the brain tissue may contribute to network stabilisation and the decline in neuronal ‘plasticity’ observed in the adult brain. However, up to date this link has not been tested directly. To close this gap, C02 will analyse the effect of unilateral enzymatic ECM removal on the mechanical properties of the brain tissue using atomic force microscopy (AFM) indentation and Brillouin microscopy (BM). In the second step, we will culture neurons in artificial ECM with variable mechanic properties but similar composition to address the effect of varying viscoelasticity on synapse formation and activity-dependent structural ‘plasticity’ assessed by monitoring the dynamic formation and growths of dendritic spines. Finally, we will increase instantaneously and reversibly the density (hence, mechanical properties) of the endogenous ECM by optogenetic means and monitor effect of these manipulation on activity-dependent spine dynamics. Taken together, these results will provide valuable information about the contribution of the ECM to brain mechanics and the role of the viscoelastic properties of the ECM in neuronal ‘plasticity’.

Project leader: Dr. Renato Frischknecht

Position: 1 doctoral researcher

C03: The role of matrix mechanics in synchronised neuronal activity’

C03 aims to use an in vitro model of epileptogenesis derived from rat hippocampus to investigate the mechano-biological aspects of neuronal circuit formation and function in mechanically tunable engineered matrices. We will focus on the interplay of structural, viscoelastic, and adhesive matrix material properties on the one hand and the formation of three-dimensional neuronal network topology, connectivity, and seizure-like hyperactivity on the other hand. The simple level of neuronal organisation in our in vitro model allows us to address whether the generation and propagation of seizure-like neural activity is directly related to matrix stiffness and viscoelasticity or indirectly caused by an altered network structure formation in response to matrix mechanics. We will then be able to test the hypothesis that mechano-chemical properties of the engineered tissue matrix can be precisely tuned to modify disease, i.e., restore normal cellular and circuit structure and function. This is aimed to provide the basis for the development of novel mechanistically informed treatment approaches in epilepsy that go beyond unspecific suppression of synaptic activity.

Project leader: Dr. rer. nat. Dr. habil. med. Katja Kobow

Position: 1 doctoral researcher

C04: Cellular differentiation in brain tissue-like matrices

C04 aims at using cell types with the same embryonic origin but distinct differentiation than brain cells, melanocytes and melanoma cells, to analyse the role of mechanics on cellular differentiation and de-differentiation. Melanocytes originate from the neural crest and display several characteristics typical of neural cells such as Schwann cells. Interestingly, melanoma cells are known to have the feasibility to transdifferentiated into Schwann cells (‘Schwannian differentiation’). Comparison of cells from melanocytic origin to neural cells with regards to the impact of the different microenvironmental matrices on differentiation is therefore highly interesting. Further, melanoma cells actively metastasise into the brain in a high percentage of patients. Understanding of this process resembles a strong clinical need as brain metastases of melanoma are still not curable. We will analyse melanocytes or melanoblast-related cells (de-differentiated melanocytes) in decellularised brain tissue or brain tissue-mimicking matrix and will use a detailed molecular characterisation (e.g., differentiation specific genes (MITF, Pax3) and senescence markers, RNASeq and bioinformatical analyses) to define the brain-tissue-specific cell behaviour and differentiation. By modelling of the matrix, its defined influence on molecular characteristics will be analysed. In addition, molecular features and tumor cell characteristics of melanoma cells in brain tissue-mimicking hydrogels or decellularised brain tissue will be studied in detail to get insight into brain metastases of this aggressive tumor. We finally aim at understanding whether differentiation cues are mediated via biomechanics and further, whether the brain is an attractive soil for melanoma cells due to its mechanical properties.

Project leader: Prof. Dr. Anja Bosserhoff

Position: 1 doctoral researcher

C05: Molecular mechanisms of neuronal mechanotransduction

C05 will develop methods to investigate molecular mechanisms of mechanosensing and mechanotransduction in primary neurons isolated from the rat hippocampus and from frog retinae. Cells will be cultured in engineered hydrogels with highly non-linear mechanical properties that allow us to spatially and temporally tune the mechanical matrix properties by imposing defined matrix strain with magnetic actuators, global stretch, and arbitrary boundary conditions in the form of flexible pillars, walls, and ridges. We will then study mechanoresponsive cell behaviour, cell morphology, and morphodynamics through time-lapse video microscopy and 3-D traction microscopy. Cell and local matrix rheology will be analysed using particle tracking microrheology. Through pharmacological and genetic perturbations of cellular force transmission pathways, we will assess potential mechanotransduction pathways. Ultimately, we will design a minimalistic ‘artificial brain’ matrix with locally controlled chemical and mechanical properties to identify a minimal set of parameters to guide axons along defined pathways. These experiments will reveal how chemical and mechanical signals play together to regulate neuronal development, and likely also regeneration.

Project leader: Prof. Dr.-Ing. Ben Fabry

Position: 1 doctoral researcher

 

Cross-Sectional Research Area X: Cross-Sectional Projects

X01: Model-based reconciliation of ex vivo and in vivo test data

X01 targets the long-standing problem of contradicting results regarding the mechanical properties of ultrasoft matter such as brain tissue when using different ex vivo and in vivo testing techniques. We aim at reconciling the results obtained from different measurement techniques within EBM including MRI, MRE, vibration tests, rheometry, indentation, and AFM. These methods individually provide mechanical, in particular viscoelastic, and biophysical parameters on different time and length scales under different physiological conditions. Our underlying hypothesis is that we can establish a continuum based model, which enables unifying the different experimentally observable regimes of the complex cerebral tissue response in vivo and ex vivo. To this end, we will substitute traditional simplifying modelling assumptions by a fully nonlinear poro-visco-elastic modelling paradigm, also enhanced by taking into account the effect of vascularisation. Using cutting-edge inverse material parameter identification by computationally simulating the various testing procedures, thereby taking into account the pertinent initial and boundary conditions, we will establish a robust identification approach to unify the information from different testing techniques. The model will be calibrated and validated with experimental data obtained in dedicated phantom materials with specific visco-elastic and poro-elastic properties, animal brain tissue specimens, and in vivo brain of mice and humans. Thus, we can for the first time use ex vivo mechanical parameters harvested from different testing modalities to explain the in vivo behaviour of the human brain. Our long-term vision is to predict disease-induced cerebral (poro-)visco-elastic property changes based on mechanical testing methods.

Project leaders: Dr. rer. nat. Jing Guo/Prof. Dr. rer. nat. Ingolf Sack, Prof. Dr.-Ing. habil. Paul Steinmann, Prof. Dr.-Ing. habil. Kai Willner

Positions: 2 doctoral researchers, 1 postdoctoral researcher

X02: Data analysis and machine learning for heterogeneous, cross-species data

Machine Learning (ML) and specifically Deep Learning (DL) has revolutionised signal and especially image processing, with unprecedented possibilities to automate quantitative analysis. While the methods developed in this field translate well across a variety of tasks, they often need large amounts of data and especially data with corresponding high-quality ground truth correspondences (gold standard measurements/labels/annotations) to adapt the parameters of the underlying models for each new application. Within the EBM consortium, a wealth of genetic, biochemical, mechanical, and imaging data will be acquired in different projects, spanning different species, experimental settings, and modalities. To utilise this data to its full extent and support quantification of experimental results, project X02 will firstly generate tools and models to integrate machine learning and deep learning techniques within the projects of this collaborative research consortium (CRC). Furthermore, we will promote and provide expertise for the use and release of collected data to online open data repositories to interface with the global research community.

Secondly, X02 will use the data acquired within EBM to investigate methods to transfer knowledge under different domain shifts, and thereby enable novel insights into measurement modalities like Brillouin microscopy and atomic force microscopy (AFM)-based nanoindentation in combination with histological analysis and fluorescence imaging. Within the consortium, in silico and in vitro data will generate significantly more specific, annotated data than in vivo experiments; similarly, more extensive measurements on organoids or less complex organisms will be available compared to higher level organisms and in particular humans. This opens the challenge and the potential of transferring knowledge and deep learning architectures trained by richly annotated, high-quality data to new domains in which data and respective ground truth correspondences are difficult to obtain. Therefore, investigating the generalizability of neural networks with the goal of improved transferability (transfer learning/few-shot learning) and easier domain adaptation will be central to this project and to the further application of machine learning within EBM. Core collaboration for this will be A01, A02 as well as B01, B02, B03 and X01. The approaches and advances developed within X02 will systematically be made available to projects in EBM that utilise machine learning models.

Lastly, informed by the aforementioned investigations, X02 will look into the modularisation of neural networks. Firstly, encapsulating low-level preprocessing and increasingly high-level analysis steps like feature accumulation and classification or regression will allow us to combine different measurements and modalities in an inherently flexible manner. Secondly, we will combine highly flexible, fully learning-based modules with known operators, which model specific physical and mechanical relationships. This will allow us to further reduce the number of parameters in the modules, thereby reducing the need for data in a new target application. Here, we will tightly collaborate especially with A01, A02 and X01 to investigate the extent to which real measurements can be complemented or replaced by machine learning-based predictions.

These three components will allow us, on the one hand, to combine insights that go beyond those of the individual projects by using machine learning. On the other hand, it further enables us to identify promising targets for machine learning approaches and the corresponding data requirements for later funding periods to integrate additional learning-based modules in the modelling frameworks developed within this consortium and improve speed, predictiveness and flexibility of these approaches for diagnosis, prognosis and treatment of neurodegenerative diseases and spinal cord injury.

Project leaders: Prof. Dr.-Ing. habil. Andreas Maier/Prof. Dr.-Ing. Katharina Breininger

Position: 1 postdoctoral researcher

X03: Engineering brain tissue-like matrices

Hydrogels are 3D networks of hydrophilic polymer chains that have been crosslinked by different mechanisms (covalent, ionic, etc.). Due to the similarities to the native extracellular matrix and the high degree of modifiability especially in stiffness and degradation kinetics, hydrogels provide ideal conditions for the use in any application where native tissue must be mimicked, replaced or simulated. X03 will engineer artificial brain tissue via hydrogel-based substitute materials with complex mechanical properties similar to those of certain areas of native brain tissue. To this end, oxidised polysaccharides (most importantly hyaluronic acid) will be combined with different proteins and extracellular matrix components to alter the material’s physico-chemical properties and to provide biochemical functionalities. The material properties will be thoroughly assessed using the different techniques used by the participating groups of EBM. To analyse the influence of specific material properties on cell fate, the amount of different chemical crosslinking functional groups will be varied to exclusively alter stiffness- or stress-relaxation behaviour while keeping the basic hydrogel components constant.

Project leader: Prof. Dr.-Ing. habil. Aldo R. Boccaccini

Position: 1 doctoral researcher

 

Central Projects

Y: Establishing magnetic resonance elastography at FAU

While several techniques enabling in vivo tissue mechanics measurements are already established at FAU, including atomic force microscopy-based time-lapse in vivo stiffness mapping and Brillouin microscopy, none of them can be applied to the in vivo human brain. The only technique currently capable of noninvasively and quantitatively mapping the viscoelastic properties of human brain in patients is magnetic resonance elastography (MRE). Furthermore, MRE is a very powerful tool that can be combined with several other MRI imaging techniques to allow a multidimensional tissue characterisation.  However, MRE-based in vivo measurements of the CNS (or other soft tissues) in patients are currently not possible in Erlangen due to the absence of the technical setup and the corresponding MRE expertise.  The aim of Project Y is to close this important gap and to establish MRE in Erlangen with the help and guidance of Jing Guo and Ingolf Sack from the Charité in Berling, who are both world-leading experts in the field of brain-MRE.

Z: Scientific coordination and fiscal administration

The misson of Z is the effective and efficient scientific coordination and fiscal administration of all centralised EBM strategic areas of action based on the dedicated centralised EBM budget. It synchronises all research activities within EBM (including research data management (RDM) and the organisation of all internal EBM executive board and member gatherings) and serves as the focal connection point for all stakeholders involved with EBM.

iRTG: Integrated Research Training Group on EBM

EBM aims to disclose the role of brain mechanics, to explore how mechanics interplays with processes during development, in health and disease, as well as during regeneration, and to engineer substitute and phantom materials. It integrates disciplines such as, e.g., experimental analyses, clinical studies, and bioengineering, all informed by advanced modelling and simulation. Consequently, the success of EBM hinges critically on establishing and exploiting interdisciplinary synergies between the EBM PLs, doctoral and postdoctoral researchers in its EBMqualify concept.

The integrated Research Training Group (iRTG) of EBM targets this demanding interdisciplinary challenge by providing a structured mandatory qualification programme and ensuring quality management and control for the doctoral and postdoctoral researchers, thereby also nurturing their scientific independence and promoting their career development. Taken together, for EBM’s doctoral and postdoctoral researchers the planned iRTG aims at:

  • Promoting cutting-edge interdisciplinary research, training, and education.
  • Providing an environment for further development of academic skills and independence.
  • Enabling effective networking of doctoral and postdoctoral researchers.
  • Developing independence, leadership, communication, and teaching skills.
  • Encouraging active involvement in EBM’s research and qualification programme.
  • Supporting development opportunities that promote academic/non-academic careers.