Designing an application to connect neural circuits and build connectomes
Role:
- Frontend engineeringBackend engineering
- UX researchUI microinteraction
2023
Switzerland
Summary
From a long, complex process, rebuilding a digital brain and modeling its electrical behavior, to a guided path to explore, build and simulate, from a single neuron to the whole brain. I ran the user research, with interviews and journey mapping, wrote the product roadmap and designed the workflows, pages and interactions. Workshops put scientists of very different brain regions on common ground, and the virtual lab brought models, simulations and explored items into one place.
Ask me
- Year
2023
- Role
Product designer: user research (interviews, journey mapping), product roadmap, workflows, pages, journeys and interactions. Frontend and backend engineer.
- Field
Neuroscience: connectomes, neural circuits, morphological-electrophysiological models
- Problem
Rebuilding a digital brain and modeling its electrical behavior is a long process, carried out by scientists who specialize in very different brain regions.
- Decision
Organize the whole application around three constant gestures at every scale, explore resources, build models, simulate them, then bring them together in a virtual lab.
- Result
One space for models, simulations and explored items, which opens the way to an AI-assisted paper-writing feature.
01
Context & business friction
A digital brain to rebuild, step by step, by experts who do not work on the same regions
Neuroscience is a vast field: each scientist specializes in one or more brain regions. The Swiss Brain Observatory had to let them connect neural circuits and build connectomes, matrices that represent the connections between thousands of neurons. The work is long and happens at several scales, from the single neuron and the ion channel up to the microcircuit and the whole brain.
02
The Key Trade-off
What I pushed through, negotiated or refused
I chose to start from workshops rather than from a single model: each team explained the problems of its daily work, and quick prototyping gave common ground between very different solutions. Through every iteration, the core concept stayed the same: explore resources, build models at different scales and simulate them. I then brought model-building experiments, simulations and the items gathered in Explore into one space, the virtual lab.
03
The Transformation
From scattered data to a guided exploration, from the brain atlas to simulation
Brain Factory is where a digital brain is configured. Its home lists the configurations that are pending, recently used and public, next to the releases and disease models they start from. Inside one, five tabs lead from the cell composition to the build: cell composition, cell model assignment, connectome definition, connection model assignment, build. The first tab shows, as a flow diagram, how the neurons of a brain region divide by layer, cell type and excitatory or inhibitory.
The connectome is defined as a matrix of the connections between brain regions, read from the presynaptic side to the postsynaptic one, at the macro, meso or micro level. A color scale gives the density of each connection, a minimap places the view in the whole matrix, and a list keeps track of every modified connection. A 3D view is one click away. Next to it, a synthesis screen runs a model on its parameters and sets its persistence barcode, diagram and image beside the neuron it produces.
In the Observatory, simulation campaigns are browsed as cards. Each one shows its configuration, status, date and user, the dimensions that vary and its attributes, with links to open it in three tools. A tab separates public campaigns from the user’s own, and a settings button chooses which columns are shown.
04
Business Impact & Proof
The value generated for research
The virtual lab brings model-building experiments, simulations and the items gathered in Explore into one place. That single space opened the way to an AI-assisted paper-writing feature.
Users keep control over the quality of their analyses and simulations, even massive ones, and an AI literature search complements the atlas to go deeper into knowledge discovery.
05
Critical hindsight & retrospective
What I learned and how mature I am
The workshops showed that each team enriches the others’ process when it exposes its real difficulties. I would repeat this format from the start, with quick prototypes in every session.
