Temporal Ensembles in Self-Aware Networks: Dendritic Recovery, Phase-Structured Recurrence, and the Hypothesis of Temporal Experience
Year: 2026
Type: preprint
Abstract: A neural rendering must connect events across time without erasing the order or ownership of their contents. I propose a Temporal Ensemble Hypothesis: receiving components whose recent activity changes their readiness can contribute in a coordinated sequence while the effects of earlier contributions remain available. A three-branch alternating image motivates the hypothesis but is not assumed to be a universal rule. I distinguish activation, recovery, retained content, coupling, downstream use, and the further claim of elementary cellular temporal experience. Under explicit hard-window assumptions, a regular schedule of n branches with recovery interval R can sustain an overlap of k distinct contributions with trace lifetime M only when n is at least k and nM is at least (k−1)R. An imposed-schedule model checks that boundary; a readiness-plus-competition toy produces a three-branch cycle under one parameter setting. A passive-I/O comparison shows that an alternating route and a one-unit ordered-history route agree on all 27 declared input sequences. A further matched-logical-intervention control agrees on all 216 input/mask cases. A fifth local-state construction has competitive activation, adaptation and decaying object-feature traces. It retains all 576 separated-event orders and all 720 supplied ownership assignments in its fixed noiseless grid. However, an ordinary nonoscillatory state bank gives the same downstream outputs: shared linear trace decay and normalized input allocation make the aggregate independent of alternation. These deliberately elementary constructions establish feasibility, actual content use in a program and identification limits, not biological oscillation, route necessity or consciousness. I propose independently localized cellular and network perturbations to test whether the candidate timing relation contributes beyond generic state dependence. Temporal computation, whole-network experience, and the proposed elementary experience of a cell remain separate conclusions. Preprint v1.0, 23 September 2026. The companion contains source, applications, frozen results, mathematical material, figures and bounded reproduction instructions. Negative and tied comparisons are preserved. Automated and AI-assisted checks are not independent scientific peer review. External raw inputs and private correspondence are not redistributed; consult the source-access and rights notices. CC BY 4.0 covers original prose, figures, documentation and results. Original software is provided for review without an added general reuse license; explicitly attributed third-party code retains its own license.
Author: Blumberg, Micah
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7214072326
Cites:
Cited by:
Related to: