Two-way tracer balance had no clear link to lower repeatable fMRI variation between ten mice in 247 region pairs

We tested whether balanced anatomical connections in both directions mark less repeatable variation in brain coupling between mice. We used an open tracer model and awake fMRI from ten other mice. Exact atlas matching gave 45 right hemisphere regions. The primary fit used 247 pairs with a detected tracer link. For each pair, we calculated the covariance across mice between Fisher-transformed coupling in two recording halves. This estimates individual coupling variance only if the half errors have zero conditional means and zero covariance. We controlled total tracer weight, distance, mean coupling, and region identity. The balance slope was +0.000535 Fisher z squared per predictor standard deviation. Its central 95% range from 4,999 mouse resamples was -0.003070 to +0.004069. A conditional node-label reference gave a two-sided fraction of 0.7595. All seven planned sensitivity ranges included zero. Added checks found that each mouse's centered coupling pattern matched its own other half. Thus the outcome contains repeated individual patterns, but two-way balance did not clearly mark less variation under the tested controls. This is not an equivalence result or a causal test. One recording per mouse, a common anatomical atlas from other animals, motion processing and vascular signals limit interpretation. We preserve compact inputs, pinned source hashes, all results and draws, code, figures, and repeat steps. A second extraction matched all compact inputs, and a second analysis matched 21 numerical and audit files exactly.

structure and function · mouse connectome · reciprocal connections · functional connectivity · individual variation · fMRI · reproducible analysis