Visual similarity after mean time course removal was linked to strength in 52 mouse V1 inhibitory connections; estimates within mice were uncertain

We tested whether the mean time course across visual stimuli explains a known relation between response similarity and synaptic strength. We used paired visual and synaptic measurements from Znamenskiy and colleagues. We split response cosine into a mean time course term and a stimulus-specific term. Both terms kept the original norms. The main panel had 52 PV-to-pyramidal connections from 18 mice. The model controlled for the mean time course term and soma distance. The specific term had a positive relation to log response magnitude. A 0.1 increase gave a fitted magnitude ratio of 1.45. Its 95% whole-mouse bootstrap range was 1.15 to 1.90. Adding the specific term reduced mean squared prediction error from 0.0937684 to 0.0816261. Each test mouse was excluded from training. The error reduction was 12.9%. A bootstrap with model refits gave a range from a 7.3% error increase to a 35.8% reduction. Thus, prediction benefit remained uncertain. An added control gave each mouse a separate baseline strength. The inhibitory ratio then fell to 1.16, with a range of 0.88 to 1.79. The data did not establish a relation within mice. Five planned inhibitory checks kept positive point coefficients. The stricter responsiveness cutoff left a range across zero. A secondary excitatory sample gave a less precise relation. The full-cosine predictor had lower error than the split model. This exploratory result shows a positive combined-sample relation after mean time course and distance controls. It does not establish a causal wiring rule. The package preserves code, source hashes, derived response arrays, results, bootstrap draws, figures, and repeat instructions. Large raw files remain at the pinned external source.

visual cortex · PV interneuron · synaptic strength · stimulus specificity · response time course · structure and function · paired recording · animal holdout · reproducible reanalysis