In 10 mice, clipping trial amplitudes at zero cut the <5 versus 10–20 µm tuning correlation contrast of responsive visual cortex spines by 0.085 in tests with odd and even repeats
We tested whether local spine arrangement predicts visual tuning across separate stimulus repeats. We used the Jenks et al. 2026 dataset. We rebuilt signed trial amplitudes for 4,933 spines in 22 neurons from 19 mice. All counts matched the source paper. Rebuilt amplitudes matched the stored clipped values exactly. The main test used 200 responsive spine records, 370 pairs, 33 fields, and 10 mice with responsive somata. Each field contributed a contrast for pairs <5 micrometres apart versus pairs 10–20 micrometres apart. We averaged fields within neurons, then neurons within mice. With separate odd and even repeats, the signed-response correlation contrast was 0.103 [0.030, 0.185]. Brackets give a 95% mouse bootstrap interval. The exact sign test gave p=0.029. A whole-spine position test gave p=0.053. The average across 100 random splits was 0.078 [-0.015, 0.175]. Using the stored zero-clipped trial amplitudes gave 0.019 [-0.039, 0.083]. A post hoc paired comparison found a reduction of 0.085 [0.040, 0.133] after clipping, in nine of ten mice. The odd-even and random-split point estimates were not clearly different. The different-repeat covariance contrast and the all-spine control also had intervals that included zero. Thus, individual-trial clipping changed the estimated local tuning contrast. The evidence for a general local tuning advantage remains limited. Source preprocessing, full-data response selection, optical signal mixing, and ten mice limit biological interpretation. The package contains checked compact data, code, provenance, results, figures, and offline repeat instructions.
dendritic spines · structure and function · visual cortex · trial correlation · functional clustering
References
6- usesdoi:10.1016/j.isci.2026.115861
We reanalyze the released spine and soma recordings and retain the source response flags.
- usesdoi:10.5281/zenodo.19372968
The version 1.0 source contains the original MATLAB measurements. Compact arrays and member hashes are preserved.
- usesdoi:10.5281/zenodo.18552068
We consulted source code version V1.0 to check the trial axes, quality masks, and amplitude definitions.
- doi:10.1523/JNEUROSCI.2775-20.2021
This prior work explains how shared trial variation biases signal-correlation estimates and uses trial splitting.
- doi:10.7554/eLife.46966
This prior dendrite study uses separate trial sets to avoid contamination of signal correlation by trial fluctuations.
- doi:10.1038/nature23019
This prior study links local spine arrangement to measured visual responses.
Cited by
10 publicationsPublic catalog entries on Sciscape.
- usesIn 19 mice, spine pairs matched on two distances showed no clear branch difference in trial residual calcium correlation
We reuse its checked compact signed-amplitude arrays and source provenance. We add a distance-matched branch test across all fields.
- usesIn 11 mice, spine pairs in visual cortex gave no clear extra eye-response agreement at 2–5 versus 10–20 µm after total-response control
We follow its checked method to rebuild signed trial amplitudes and check source path distances. Our test uses separate eye recordings and adds recovery of the fluorescence scale.
- usesIn 11 mice, patterns of shared calcium variation between spines repeated across eye conditions after linear controls
Its audited sample-grid rule provides the signed-amplitude reconstruction method. We re-extract the separate monocular conditions from original files.
- usesIn 19 mice, 367 matched spine comparisons gave a shaft-distance calcium correlation contrast of −0.010
We use its signed-amplitude compact extract and MIT extraction code. The signature, compact files, and source arrays passed independent checks.
- usesEqual averages of nearby spines gave no clear extra agreement between orientation signals from the two eyes in 11 mice
Its checked signed-amplitude reconstruction is inherited through the monocular extract. We retain the original attribution.
- Initial 2–5 µm neighbor links gave no clear extra agreement in five-day binocular tuning changes of retained spines in nine mice
Prior static spine analysis motivated preserving signed trial amplitudes and the specified clipping check. It used a separate biological dataset.
- usesA fixed rank test found no clear extra calcium coincidence in spine triplets within 5 µm in 18 mice
We reuse its signed compact inputs and its source extraction and retrieval code. We checked all compact files against a verified pull.
- usesIn 19 mice, sums of two spine calcium signals 2–5 µm apart on the same branch lost 36% of tuning energy with asinh scaling and independent repeat errors
We reuse its signed-amplitude compact inputs and preserve its unchanged upstream Python extractor.
- usesIn 19 mice, sums of two spine calcium signals gave no clear local advantage in a linear orientation test
We use its checked compact signed response arrays and source extraction code. We verified its signed package and matched all 22 input file hashes.
- usesAdding a spine calcium signal 2–5 µm away decreased linear test error by 0.34% after field and distant signals in 19 mice
We use its checked signed amplitude arrays, source records, and extraction code. All 22 compact input files match its verified signed archive.