STIMULUS / ART FEED
01 / 03
REFIK ANADOL STUDIO
Gallery B
Dataland
00:0000:00
BRAIN / SOMA ATLAS
MaleCNS v1.0
AnatomyVisual inputFrame change
VIDEO-DRIVEN OVERLAYIllustrative · not neural activity
124,289 measured somataData & credit
FLY / PHONE INPUT
Flybody Same video · same time
Anatomical body modelNext in 5.5s
STIMULUS TIME0.00 s
VIDEO01 / 3
FRAME CHANGE0.0 %
IMAGE LUMINANCE0.0% · visual input
*Visual input preview: optic-lobe color follows video brightness and frame changes. This illustrative overlay is not measured or predicted neural activity.
EXPERIMENT NOTES / 001
What does a fly see in art?
A visual stimulus workbench. Scroll through Refik Anadol Studio clips while the fly watches the same feed on its phone. The atlas displays an explicitly illustrative visual-input overlay while a neural model is pending.
Data provenance · measured anatomy, illustrative overlay
Each point is one measured cell-body location from the adult male Drosophila MaleCNS v1.0 dataset. We display 124,289 classified optic, central and descending somata. VNC-associated and unclassified cells are omitted from this view; 140,024 measured positions remain in the downloadable atlas. This is a soma map, not a rendering of neuron branches or synaptic connections.
Native coordinates use isotropic 8 nm voxels. The display only centers, rotates and uniformly scales them. No generated positions, axis stretching or smoothing. “XY view” returns to a fixed native-coordinate projection for comparison.
Blue is anatomical context. Cyan and white visualize image brightness and frame changes; the image-to-brain placement is illustrative, with no biological receptive-field mapping. No measured firing, trained model output or connectivity simulation is shown.
Data credit: FlyEM / HHMI Janelia, University of Cambridge, MRC Laboratory of Molecular Biology and Google Research. CC BY 4.0. Adaptations: annotation filtering and binary export; display rotation, scaling and coloring. No endorsement implied.
Source, filters and checksums · Data notice · Project and scientific publication
THIS SESSION
Three videos share one playback clock. The fly scrolls every 5.5 seconds while playing; you can also scroll manually or turn auto scroll off.
MODEL LATER
Video frames will pass through a fly visual encoder into a connectome model. The current overlay uses an 8×4 image grid of brightness and pixel changes, without a retinotopic or neuron-level mapping. Actual activity must reference the same neuron IDs and stimulus time. Human TRIBE brain predictions are not fly activity.