BRAINHACK DONOSTIA · OPEN SCIENCE

Your research toolkit.

Find the tools to code, collaborate, manage data, review research, and share what you build. From your first project to your next reproducible pipeline.

CORE TOOLKIT

General tools

Tools that help everyone code, organize, share, and document their work.

GitHub

Version control · Project collaboration

Free tier · Proprietary platform

CodingCollaborationProject managementReviewReproducibility

What is it?

GitHub is a collaborative platform built around Git, a distributed version control system. It lets teams host repositories, preserve the history of changes, and develop different parts of a project in parallel. Beyond storing code, GitHub brings discussions and decisions closer to the files they concern: Issues record tasks and problems, Pull Requests make changes reviewable, and Projects offer flexible ways to organize work. Repositories can be private while a project is taking shape or public when an appropriate sharing plan is in place.

Why use it at BrainHack?

During BrainHack, contributors can work on the same research project without passing competing versions of files back and forth. Beginners can read the documentation, report problems, or take on a clearly defined Issue; experienced developers can implement changes and review Pull Requests. A well-maintained repository creates a useful record of how a project evolved and helps work continue after the event. Git is open source, but GitHub itself is a proprietary hosted platform—an important distinction when discussing open science.

VS Code & VSCodium

Extensible code editors

Free editors · Different licensing

CodingCollaborationDocumentation

What is it?

Visual Studio Code and VSCodium offer similar editing experiences for programming, scripts, configuration files, and Markdown documentation. Both support syntax highlighting, debugging, integrated terminals, source control, and extensions for scientific workflows such as Python, R, and Jupyter notebooks. Visual Studio Code is Microsoft's distribution of an editor whose underlying source code is openly available. VSCodium is a community-built distribution of that open codebase, designed without Microsoft's built-in telemetry and using an alternative extension registry.

Why use it at BrainHack?

BrainHack participants can use either editor to inspect an unfamiliar project, edit scripts, compare file changes, and commit their work. The extension ecosystem makes these editors useful across multiple research fields, from experimental software to neural-network models. Offering both options illustrates a broader open-science lesson: software may be free to download without its distributed product being fully open source. Some proprietary extensions are easier to obtain in Microsoft's version, so participants should choose the editor that best fits their project's requirements.

JupyterLab

Interactive scientific computing

Free · Open source

CodingCollaborationReproducibilityData management

What is it?

JupyterLab is an interactive environment for notebooks, code, data files, and scientific visualizations. Its notebooks combine executable cells with explanatory text, mathematical notation, tables, and figures in a single document. Researchers can inspect intermediate results rather than running an entire analysis as one opaque program. Python is common, but other languages, including R, are supported through separate kernels. JupyterLab can run on a participant's computer or through a hosted server; real-time editing is possible when collaboration has been explicitly configured.

Why use it at BrainHack?

The notebook format is a friendly entry point for participants with different levels of experience. A beginner can change a parameter or rerun a plot, while an experienced collaborator develops the underlying analysis. BrainHack teams can explore EEG or fNIRS signals, visualize model behaviour, and explain results step by step. Notebooks become especially useful for sharing methods when their dependencies, inputs, and execution order are documented. Real-time collaboration is not automatic on every installation, and notebooks should still be versioned carefully when edited through Git.

Docker

Portable research environments

Docker Engine: open source · Desktop terms vary

CodingReproducibilityCollaboration

What is it?

Docker uses containers to package applications together with the software libraries and system dependencies needed to run them. Instead of asking every collaborator to recreate an environment from separate installation instructions, a project can describe the environment in a Dockerfile or distribute a prepared container image. A container is not a full virtual machine, but it can make software more portable and predictable across compatible systems. Docker Engine includes open-source components; Docker Desktop has separate licensing conditions that participants should check.

Why use it at BrainHack?

Scientific workflows can fail simply because two computers use different software versions. Docker reduces this friction and makes it easier to rerun a pipeline after a hackathon has ended. For BrainHack, a shared container can help participants launch the same analysis tools or machine-learning dependencies without manually resolving every package conflict. Containers support computational reproducibility, but they do not replace proper documentation, versioned data, or controls over sensitive research materials. GPU-intensive projects also need compatible hardware and container configuration.

Zotero

Research references & shared libraries

Free · Open source

DocumentationCollaborationSharing

What is it?

Zotero is an open-source reference manager for collecting, organizing, annotating, and citing scientific literature. It can capture bibliographic information from journal websites, store attached PDFs, and help users keep references consistent across a project. Its word-processor integrations simplify inserting citations and producing bibliographies. Group libraries provide a shared place to organize papers and references with collaborators. Zotero is a research organization tool, rather than a code editor or a system for versioning scientific datasets.

Why use it at BrainHack?

BrainHack teams often need to understand a method or software package quickly before changing it. A shared Zotero library can assemble the core papers behind a project's approach, preserve useful annotations, and make it easier for newcomers to enter the discussion. It is particularly valuable for people contributing theoretical, methodological, or literature-review expertise rather than code. Because the application is open source, its implementation is inspectable and can be extended by the community. Availability of cloud storage for attachments depends on the team's storage arrangement.

PROJECT-SPECIFIC TOOLKIT

Beyond Speech

Neurophysiology workflows, scientific data management, and open standards.

∿

MNE-Python & MNE-NIRS

EEG & fNIRS analysis

Free · Open source

Specialized researchCodingReproducibilityData management

What is it?

MNE-Python is a scientific Python toolkit for processing, analysing, and visualizing neurophysiological recordings, including EEG and MEG data. It provides established methods for filtering, event handling, artefact inspection, epoching, and statistical analysis. MNE-NIRS extends this ecosystem with tools tailored to functional near-infrared spectroscopy, including haemodynamic signal processing and modelling. Together they provide a scriptable analysis environment in which steps can be inspected, adjusted, and documented instead of remaining hidden in manual software operations.

Why use it at BrainHack?

Beyond Speech studies how sign-language immersion relates to language-associated brain activity and proposes an open EEG/fNIRS workflow. MNE-Python and MNE-NIRS are therefore directly aligned with the project's data modalities. Participants can examine preprocessing choices, compare visualizations, and develop reusable analysis scripts under version control. Pairing the two tools in one card avoids separating closely related components of the same scientific workflow. Their open-source development also supports methodological transparency: researchers can inspect implementations, report issues, and contribute improvements.

DL

DataLad

Versioned research data management

Free · Open source

Data managementSharingReproducibilitySpecialized research

What is it?

DataLad is a tool for managing datasets and the relationship between data, code, and computation. Built around Git and git-annex, it allows datasets to be versioned while large files can be stored and retrieved separately from lightweight metadata. Researchers can obtain the structure of a dataset without necessarily downloading every recording. DataLad also supports workflows that record how analyses were run, helping link outputs back to their inputs and processing steps. It is particularly relevant when datasets are too large or complex for ordinary Git usage.

Why use it at BrainHack?

Beyond Speech may involve large electrophysiological and optical recordings shared among researchers. DataLad can help the team maintain an organized, traceable dataset while controlling where larger files are stored. This is useful during BrainHack when some participants need the full recordings and others only need code, documentation, or small examples. The approach can support reproducibility after the event, but access permissions and participant confidentiality still need separate safeguards. A DataLad dataset being technically shareable does not mean the underlying human-subject recordings may be publicly released.

B

BIDS & MNE-BIDS

Open standards for neuroimaging data

BIDS: open standard · MNE-BIDS: open source

Data managementReproducibilitySpecialized researchSharing

What is it?

BIDS, the Brain Imaging Data Structure, is a community-developed standard that describes how research data and metadata should be organized. It uses predictable directory structures, filenames, and accompanying descriptions so that datasets can be understood beyond the laboratory that collected them. MNE-BIDS is a Python package that connects MNE analysis workflows with BIDS-formatted data, including supported EEG and fNIRS recordings. BIDS is a standard rather than an application; MNE-BIDS is one of the software tools used to work with that standard.

Why use it at BrainHack?

A common format is vital when a BrainHack team brings together people using different techniques and software. For Beyond Speech, BIDS helps ensure EEG and fNIRS recordings are arranged consistently and accompanied by necessary metadata. Participants can validate dataset organization, understand what each file represents, and write analysis pipelines that are easier to reuse. Because BIDS is an open community standard, it supports interoperability instead of tying a dataset to one proprietary program. It does not by itself authorize sharing identifiable recordings or resolve research-ethics requirements.

PROJECT-SPECIFIC TOOLKIT

Reproducibility Checks

Tools for transparent scientific review and repeatable verification.

GitHub Actions, Issues & Pull Requests

Transparent review & automated checks

GitHub platform tools · Mixed licensing

ReviewReproducibilityCollaborationProject management

What is it?

GitHub Issues, Pull Requests, and Actions form a practical workflow for collaborative verification. Issues record what needs checking, including missing information, unexpected results, or requests for clarification. Pull Requests show proposed changes and give reviewers a place to discuss code and documentation before accepting them. GitHub Actions can automatically execute predefined tests, run scripts, and report outcomes when files change. These features serve different purposes, but together they connect scientific questions, documented decisions, and computational checks in one traceable project history.

Why use it at BrainHack?

The Reproducibility Checks on GitHub project explores how research verification can become more open and collaborative. Team members could use an Issue to describe a claim or figure to verify, a Pull Request to propose a correction or report, and an Action to run repeatable computational tests. This makes responsibilities and evidence easier to inspect. Automated tests verify only the conditions they were designed to check, however; passing checks does not establish the full validity or independent reproducibility of a publication. GitHub's hosted service is proprietary even when repository scripts are open source.

PROJECT-SPECIFIC TOOLKIT

RNN Night Fever

Building and exploring recurrent neural-network models.

PyTorch

Neural networks & computational modelling

Free · Open source

CodingSpecialized researchReproducibility

What is it?

PyTorch is an open-source framework for numerical computing and deep learning. It provides tensors, automatic differentiation, GPU acceleration where supported, and tools for building and training neural networks. Researchers can create recurrent architectures, define custom learning rules, and inspect internal activations throughout training. Because PyTorch is code-driven and integrates with the broader Python scientific ecosystem, it is well suited to exploratory modelling where the network structure or experiment may change repeatedly. Reproducible results still require controlled random seeds, software versions, and documented settings.

Why use it at BrainHack?

RNN Night Fever investigates whether replay-like activity can emerge from interactions among recurrent networks representing different components of a memory system. PyTorch offers the flexibility needed to build and compare these architectures, simulate learning and offline phases, and visualize how activity evolves. Beginners may begin from a prepared notebook; experienced contributors can refine training procedures or investigate model dynamics. PyTorch is our recommended implementation framework, not a technology explicitly required by the project's published description. Its open-source status also helps teams share and inspect modelling methods.

PROJECT-SPECIFIC TOOLKIT

Brainbow

Training models to detect neurons in microscopy images.

BIO

BiaPy

Deep learning for biological images

Free · Open source

Specialized researchCodingData managementReproducibility

What is it?

BiaPy is an open-source deep-learning framework focused on biological image analysis. It provides workflows for image segmentation, object detection, and related tasks in two-dimensional or three-dimensional data. Researchers can configure training and prediction pipelines, work with annotated images, and evaluate model performance. Compared with writing every part of a computer-vision pipeline from scratch, a specialized framework can provide reusable processing steps and established training workflows. The quality of the results still depends strongly on annotation quality, representative training data, and appropriate validation.

Why use it at BrainHack?

The Brainbow project aims to detect neurons in colourful microscopy images by creating annotations and using them to train a model. BiaPy is specifically named in the project plan, making it an especially strong fit for the toolbox. Team members can contribute by preparing labelled images, checking predictions, refining the training configuration, or measuring how reliably cells are detected. Publicly releasing annotations and model information can make the outcome useful beyond BrainHack, subject to the dataset's licence and research requirements. Training larger models may benefit from a compatible GPU.

Bonus tools · Further resources (4)

Optional resources for more specialized reproducibility workflows. These are not required tools for the BrainHack projects.

BrainHack Donostia Toolbox · Descriptions are guidance, not project installation requirements. Check each project's instructions before installing software.