Explore Projects

Virtual Reality for Distributed Research (VRDR) encompasses a VR showroom, immersive anatomy exploration, model organism studies, and a VR Morris Water Maze. Additionally, there are plans to enhance VR anatomy and organism models by integrating high-fidelity 3D representations with detailed specifications for a more realistic and informative experience and the development of a VR AI navmesh for the Morris Water Maze, incorporating advanced spatial mapping and artificial intelligence. An AR VR application which is compatible with all the devices such as Meta quest ,Web XR . This will in turn allow learners and researchers from around the world to access and interact with the models. This will also incorporate haptic feedback for users to make it more interactive.

Building recommendation engine for suggesting reviewers using Natural Language Processing and Huggingface library. I will be using the editors’ abstract of the data set. Apply pre-trained Bert embedding. After I have generated encodings for all the editor’s abstract present in the data set, I need to create encodings for the words of interest from the reviewer’s information extracted and find similarity between reviewers’ research interests and the encodings of the editor’s abstract. I will cosine similarity or other similarity formulas to determine the similarity between the vectors. Higher cosine score signifies the more similarity between the two vectors. We can then query the data set using numerous reviewers’ interests and rank the cosine similarity scores along with their corresponding editors’ abstract. Deliverables: I plan to accomplish the following over the summer: - Apply BERT models from HuggingFace on editors abstract for topic modeling. - Then find similarity scores between different topics using cosine similarity. - Serve the model on the web. - Documenting, publishing on Jekyll pages.

Gesture-based music generation has existed for some years now, thanks to software such as Wekinator. However, they use machine learning methods that have been overtaken by recent Deep Learning innovations, and therefore limit the studies that can be carried out on musical creativity and sound/movement interaction. The aim of this project is to use these modern Deep Learning methods to generate music driven by motion capture. The framework consists of 3 distinct modules. Firstly, user gestures are detected using landmarks generated by MediaPipe by Google. Secondly, these landmarks are used to recognize the gesture performed by the user, and generate Open Sound Control (OSC) signals. Finally, the OSC signals are converted into music using Max/MSP. In addition, new gestures can be dynamically added to the framework, with just a few examples, and mapped to new sounds. Deliverables: - Experiment to estimate the appropriate delay between a gesture and the resulting sound/effect. - Framework based on LivePose or Pose2Art using only the gestures predifined in MediaPipe, able to interface the chosen camera and the sound-card - Model based on MediaPipe holistic model to classifiy dynamic gestures - Famework to map gestures to sounds through OSC signals. - Few-Shot Learning-based model to dynamically add new gestures

This project studies the development of a tiny worm called Caenorhabditis elegans, focusing on how its cells divide, grow, and interact during its early embryonic stages. To map out and understand these complex interactions, the project upgrades the OpenWorm DevoGraph library using an advanced deep learning tool called the Explainable Spatio-Temporal Graph Evolution Learning (ESTGEL) model. This model tracks two critical dimensions of growth simultaneously: the 3D spatial arrangement of cells (how they physically touch and communicate in space) and the directed cell lineage (the family tree detailing how cells evolve to form different tissues). By combining these spatial and lineage pathways with an "edge attention" mechanism, the tool highlights the most critical cell-to-cell connections at various stages of growth. Ultimately, the project aims to get a complete, computationally explainable picture of the worm's development, from a single cell to a fully formed organism, and to successfully classify between healthy and mutant embryos. This work will help us learn more about basic life processes, track how early cell interactions influence neural wiring, and improve fields reliant on understanding cellular growth. The project is supported by the OpenWorm community, an international organization dedicated to using advanced dynamic graph neural networks to model and study the brain and nervous system development of the first virtual organism in a computer.

<p><code>imjs</code> and <code>im-tables</code>, written in CoffeeScript are client-side libraries for querying mine instances and displaying data in tabular format respectively, which help in reducing boilerplate code for developers, looking to work with InterMine's services. The aim of this project would be overhauling these libraries, by upgrading their dependencies (last updated in 2015) and adding easily extensible mock responses (in <code>imjs</code>). Moreover, the test suite will be broken up into unit and integration tests. Support for querying with the Registry class will be added. Docs for both libraries will be added (adding end-user docs for <code>imjs</code> and developer docs for im-tables). Build system of both the libraries will be investigated, and the possibility of updating it will be discussed and implemented (possibly in the GSoC period itself, if feasible as per investigation). High priority issues, like adding support for <code>setConstraintLogic()</code> in <code>imjs</code> will also be resolved.</p>

We aim to further enhance model development with NetPyNE’s “batch” subpackage by refactoring the code base for ease of use and scalability, then use the improved batch subpackage to explore and typify the effectiveness of various search algorithms (random, population based, various posterior based…) on a diverse model repository including: rodent motor(M1), rodent somatosensory(S1), and macaque auditory (A1) thalamocortical circuits to make NetPyNE’s capabilities more efficient for computational neuroscience research.

During Google Summer of Code 2022, I propose a project in which these feedback loops (recurrent relationships) and causal loops (reciprocal causality) are thoroughly examined underlying the requisites of the ethical regulator theorem. By using NetLogo to construct an analytical model that can be used to incorporate these different types of loops, a homeostatic system will be constructed that can be used to encourage cooperative and altruistic interactions - with simulated data being generated through an agent-based model of open-source behaviors and interactions. At the end of this project, it should be known, it shall be known whether the scientist (myself) can truly “know,” as Heinz von Förster stated.

To participate in the Neurobagel query federation, datasets must conform to Neurobagel’s data model, so annotating the datasets is necessary to harmonize them for query federation. The project aims to reduce the human effort to manually annotate individual data elements by automating the current annotation tool provided by Neurobagel using Large Language Models (LLMs). The already existing annotation tool will be integrated with an LLM-based assistant which will categorize and annotate each data element and it will be followed by human verification. The project includes automating the tool using LLMs, then integrating the tool into the existing webpage and making changes in the UI accordingly.

Cerebral cortex is the outermost layer of the brain and is associated with highest mental capabilities like consciousness, emotion, reasoning, language and memory. It is concerned with most of cognition and behavior so it becomes important to have accurate cortical models. This project will study already published models and convert to NeuroML/PyNN and make it available to the community through Open Source Brain. Deliverable: 1) Bi-Weekly/Monthly updates in powerpoint or word format during the work period. 2) Final report and frozen source code at the end of project. 3) User manual and instructions on using the source code. 4) Source code will be made available on github repository - OpenSourceBrain .

Eye tracking can be used for a range of purposes, from improving accessibility for people with disabilities to improving driver safety. However, modern state-of-the-art mobile eye trackers are costly, often bulky devices that require careful setup and calibration, and they tend to be expensive. In this project, therefore, we aim to develop an affordable and open-source alternative to these Eye Trackers. In this project, we improve upon the basic implementation of Google's paper which is able to obtain very good performance using simple CNNs. We implement the architecture and train the model, run SVR experiments, and compare the results with Google's own implementation. We also need to connect them with an android application for data collection, as well as for evaluating the model’s performance in real time.

The project is about incorporating new features into an existing 3D viewer which allows for visualisation of cells and networks using the NeuroML standard. Right now the viewer is limited to only viewing the model, whether it is a single cell or a complete network, but at the end of the project more options regarding the visualization should be added. The user should be able to get as many information as he can about the model using an intuitive GUI with a short learning curve, using “on click” or “on selection” methods to provide according information and providing easy access to views of cells showing ion channel distributions, segment groups, etc. . Options for using the Vispy viewer inside Jupyter notebooks/JupyterLab are also a crucial part, as Jupyter notebooks are gaining popularity rapidly.

<p>The Workflow Designer is a prototype web-based application allowing drag-and-drop creating, editing, and running workflows from a predefined library of methods. Moreover, any workflow can be exported or imported in JSON format to ensure reusability and local execution of exported JSON configurations. The application is primarily focused on electroencephalographic signal processing and deep learning workflows.</p> <p>Currently, the entire Workflow Designer system (workflow system, deep learning models, server) is based on Java. The aim of this project is to transfer the deep learning workflow related blocks/models and backend technologies from Java to Python and allow executing workflow blocks (methods) implemented in Python, using e.g. MNE for EEG signal processing, or TensorFlow for deep learning.</p>

​The Human Neocortical Neurosolver (HNN) is an open-source neural modeling tool designed to help researchers and clinicians interpret human brain imaging data. Currently, HNN primarily focuses on simulating neural activity within single neocortical columns. This project aims to extend HNN's capabilities by developing a Python API that facilitates multi-network simulations, enabling the modeling of interactions between multiple neocortical networks. This enhancement will provide neuroscientists with a more comprehensive tool to study complex brain interactions and better understand the neural mechanisms underlying MEG/EEG data. The project will involve designing and implementing the API, developing comprehensive documentation and tutorials, ensuring compatibility with existing HNN functionalities, and conducting validation tests to verify the accuracy and reliability of multi-network simulations. By expanding HNN's capabilities, this project will support a broader range of research applications, fostering innovation and collaboration within the neuroscience community

TheVirtualBrain (TVB) is the first integrative neuroinformatics platform for the modeling of full brain network dynamics. TVB simulator, written in Python, enables the systematic, model-based inference of neurophysiological mechanisms on different brain scales that underlie the generation of macroscopic, commonly used neuroimaging signals (EEG, MEG, fMRI). In the tvb-ecosystem, there is a new repository called tvb-widgets, which consists of UI widgets for Jupyter notebooks. Offering TVB-web interface's functionality as a JupyterLab widget will enhance integration, workflow efficiency, and interactivity making it a valuable addition to the TVB ecosystem for research and educational purposes. The project focuses on enhancing the tvb-widgets repository and developing new widgets. The proposal is for the implementation of more features from the Connectivity cockpit of TVB's web interface as interactive widgets primarily the Space-time Visualizer widget and the Connectivity Matrix Editor widget. This can be achieved by leveraging different libraries like Ipywidgets, PythreeJs, etc. Essential deliverables include classes, functions, tests, etc., for these widgets.

The annotation of research data is essential to ensure its findability and reusability. High-quality data annotations require domain expertise, so that annotations are relevant, but also specific technical skills, such as knowing how to handle JSON/XML files. Additionally, people are often reluctant to change their workflows, and the technical affordances in the case of data annotation intensify this challenge. As a result, researchers tend to stick to their “data handling traditions” as soon as data operations become too complex. Unfortunately, this often means that even though projects like Neurobagel are working hard to make life easier for researchers, these tools are not widely adopted. My idea for contributing to the Neurobagel project is to combine a user-friendly interface with a Large Language Model (LLM) approach to make the annotation of tabular data even more effortless for researchers. For the end user, the process should be to provide a tabular file and get a first-pass annotation for review without any intermediate steps. From a technical perspective, this should be accomplished by using a Large Language Model (LLM) to categorize columns, such as participant ID or age. To improve the predictions of the LLM, already annotated data will be linked to explicit knowledge from existing ontologies such as SNOMED CT or the Cognitive Atlas and used to provide context for the LLM.

Neuroptimus, an open-source parameter optimization software tailored for neuroscience applications, has been instrumental in advancing the construction and optimization of biologically detailed models, utilizing algorithms such as evolutionary algorithms and swarm intelligence. Neuroptimus includes a graphical user interface (GUI), and works on multiple platforms including PCs and supercomputers. Building upon its success, this proposal aims to enhance the GUI of Neuroptimus to further streamline the parameter optimization process.The project targets to address key areas of improvement within both the graphical user interface and command-line interface of Neuroptimus. By integrating missing functionalities and refining existing features, the goal is to enhance user experience, flexibility, and accessibility. Specific enhancements include enabling users to save and load optimization settings seamlessly within the GUI, providing real-time monitoring of optimization progress, enhancing the interpretability of optimization results through various output formats, and facilitating analysis and visualization of both intermediate and final optimization outcomes. Through these enhancements, the Neuroptimus GUI will become more informative, user-friendly, and conducive to efficient parameter optimization for biophysical models in neuroscience research. Project Deliverables: 1- Implementing Saving and Loading Optimization Settings. 2- Implementing Progress Monitoring Functionality. 3- Implementing Saving Optimization Results in an Interpretable Format. 4- Extending UI with Visualization Options for Final and Intermediate Results. 5- Researching and Proposing Additional UI/UX Enhancements.

<p>Fingerprinting for Programs is aimed at analyzing code blocks on a semantic level. This is done by symbolically executing the code block via SPF over JPF in order to generate path constraints for each possible path that a program may take. The path constraints are then put into canonical form and analyzed. Level of refinement of the path constraints dictate the level at which we model programs i.e. If we do not refine the path constraints at all, the program is modeled on a syntactical level. If we stop refinement early, the program is modeled on a semantic level but syntactical structure of the program is still considered. If we completely refine the path constraints and put them into canonical form, the program is modeled on a semantic level with no consideration towards the syntactical structure of the program. This gives us options on applications of fingerprinting for programs. Further research in this field could result in semantic patterns not just being observed but rather analyzed. Applications of this analysis include the possible creation of a semantic auto-complete for programmers.</p>

<p>Building a library that enables conversion of Tensorflow deep neural network models to corresponding Spiking Neural Network architectures, with minimal performance losses.</p>

<p>JPF is the most popular model checking tool for Java applications. It is extensible and there are lots of extensions for various purposes. Jpf­nhandler is one of such extensions. Its goal is delegating the execution of SUT methods from JPF to JVM level. One goal of this project is to improve jpf-­nhandler performance using the cache layer. The latter prevents invoking methods more than once and problems caused by this. It also considerably increases the jpf­nhnandler efficiency as each method is executed only once. If jpf­nhandler encounters delegated method that has been executed, it just reflects result of its invocation. Another goal is to extend the converter component and including convertor classes for missing incompatible model classes.</p>

I will seek to create GNNs that resemble actual biological networks found throughout development by analyzing time-series microscopy data. This project will provide a scaffold for modeling biological networks with GNNs, which will help to simulate the development and test various theories about biological networks.

<p>The goal of this project is to create a platform that allows the user to run simulations with Contextual Geometric Structures. This model allows for us to view evolutionary biological simulations in a way that has never been done before. Once this is done I will create benchmark tests for performance and then create a predictive algorithm that predicts the outcome certain cultures will have over time.</p>

I am proposing a project for Google Summer of Code to develop an interactive SWC to NeuroML converter in Python. The converter will simplify the conversion of the widely used SWC format to the NeuroML standard for biophysically detailed neuronal models. The project will include an interactive tool for users to utilise, and it will be tested against neuronal reconstructions from NeuroMorpho.org

This project will have the aim to derive the empirical HRF for each brain region and each subject from BOLD recordings, and then to use this HRF in a computational model of brain activity (The Virtual Brain, implemented in EBRAINS). The legacy approach will be compared with the proposed one.The final pipeline will be implemented and validated on EBRAINS, ensuring full compatibility with the TVB cloud environment.