REANALYSIS · MOUSE BRAIN · V3

Reanalyzing how the mouse brain encodes task and movement information.

Sage-A1.5 reanalyzes existing Findling2025 / IBL data from 201 mouse brain regions through SeekBrain, revisiting the shared-axis analysis reported in the SeekBrain paper. [1]

REANALYSIS RESULT

Across the mouse brain, task encoding showsa shared organization associated with wheel movement.

KEY EVIDENCE

Enter the research workspace ↗
FIVE MEASURES. ONE PRINCIPAL AXIS.Explore a brain region
Visual stimulusWheel choiceFeedbackWheel speedSigned velocity
PRINCIPAL AXIS ACROSS FIVE MEASURES—of total variance in standardized measures
MOpPrimary motor area
PC1 +2.29
201 mouse brain regionsIBL / ALLEN ATLAS
MANUALLY ORCHESTRATED

Researchers assemble each step

Prepare data, write analyses, run statistical checks, and map results onto anatomy.

AGENT-LED RESEARCH / Sage-A1.5

Set the goal. Connect the workflow.

Connect data, tools and analysis in SeekBrain to deliver results that can be inspected. [1]

SCIENTIFIC VALUE

Locate leads for the next experiment

Connect statistical relationships to specific brain regions, providing a basis for testable hypotheses.

[1] Jiamin Wu et al. (2026). SeekBrain: An Autonomous Multi-Agent System for Accelerating Neuroscience Discovery. arXiv:2607.29347.

01 / FROM QUESTION TO EXECUTION

One workspace.
A connected investigation.

Read data. Write analyses. Test relationships. Map anatomy.
See an agent turn a research objective into inspectable results.

Sage-A1.5 · V3 / REANALYSIS IN MOTION
00:00 / 00:28

Workflow reconstruction from real artifacts · The question is a task summary, not an execution transcript

SEEK BRAIN. FIND CONNECTIONS.

A researcher sets the objective. The agent connects tools and data to address the scientific question.

02 / REANALYSIS RESULTS

Different information.
A shared axis of variation.

Five task and movement measures covary across regions, suggesting shared organization. This statistical axis is associated with wheel movement.

Existing data. Leads for the next investigation.

Move from correlated measures to regions worth investigating. This agent-led reanalysis of existing data identifies regional associations for future experimental testing.

Explore more showcases ↗