The legal landscape behind the Toolkit

How 93 countries in Africa, Europe, and Latin America and the Caribbean treat the transfer of brain data, and what that means for a collaboration between any two of them. Legal experts in each region reviewed 477 laws and regulations across six layers of governance, and every one of the 8,556 country-to-country pathways is rated for the conditions it imposes in practice. Hover for detail, click to pin.

Six layers of governance

We look at each jurisdiction through six layers, from how it recognises neural data to what its contracts must carry. The layers build on each other: what a jurisdiction recognises (layer 1) determines which lawful bases are available (layer 2), which determines whether and how transfer can occur (layer 3), and only then does the question become whether protection remains effective in practice (layers 4–6). Click a layer.

The compatibility matrix: every pathway between every pair

Rows are exporting jurisdictions, columns receiving jurisdictions; each cell is the class assigned to that direction. A1 (dark green) is a unified legal space, A2 (light green) highly convergent architectures, B (yellow) pathways supported by standard transfer instruments, C (orange) operationally conditional pathways, D (red) pathways restricted in practice and E (grey) insufficient legal certainty. Hover a cell for its class; click a cell to pin the pair and read both directions and both countries; click the diagonal for a single country's profile. The classes are structured indicators of practical navigability, not findings of legal adequacy.

Where sharing is enabled

The network keeps only pathways of class A1, A2 or B (weighted 3, 2, 1). Node size is the sum of weighted in- and out-strength; the European cluster is dense because every EU pair is A1 and the UK and Switzerland are A2 with all of them. Click a node to show its enabled pathways; the thin lines are B, the thick ones A1/A2.

On the map: receiving data from Europe, or from anywhere

Each jurisdiction is shaded by the modal class of the pathways that originate in the sampled European jurisdictions (24 origins; 23 for European destinations). Outside Europe the dominant condition is C: transfers are possible, but practical governance depends on supplementary contractual, institutional or repository controls. The map is interactive: hover a country for its value, click it to pin its profile, or change the origin to see the conditions any sampled jurisdiction faces when exporting to the rest of the sample. The United States and Canada are shown in dark grey because they were analysed as governance units rather than at country level.

On the map: how varied the conditions are

Jurisdictions are shaded by the Shannon entropy (bits) of their incoming compatibility profile across the six classes. High entropy means a researcher there faces very different conditions depending on where data come from; low entropy means uniform conditions, whether uniformly enabled or uniformly restricted, so entropy is read together with the classes themselves. European jurisdictions score highest because they receive A1 from EU members, A2 from the UK and Switzerland, and B or C from the other regions.

Region to region

A summary derived from the matrix: the share of each class for every origin-region to destination-region corridor. Europe to Latin America is predominantly C (80.8%) while the reverse direction clusters in B (76.9%), an asymmetry worth knowing before you plan a collaboration; Europe to Africa is C for 65.1% of pathways.

From the landscape to a decision

The Toolkit turns this landscape into a decision for your collaboration: classify the data, look up the pathway, reassess derivatives and models, select the agreement modules. See the workflow or open the Toolkit.

Data, code and citation

The matrix and the code behind these figures are open: github.com/francopestilli/data-governance-access-matrix. Legal content reflects instruments in force at 31 December 2025. If you use the data, please cite the study: Botes M, Mendes A, Huffman B, et al. A decision support toolkit for AI-ready neuroscience data governance (2026); full citations are on the About page. The regional distributions of governance attributes for each layer will be published with the legal landscape handbook.