arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2610.19949

Reciprocity coordinates a three-way trade-off in recurrent neural circuits

Xiaoyu Zhang·Mingyi Huang·Yixuan Cui·Xin Lu·Yuguo Yu
View original source ↗Read source document ↗Export metadata ↓

Abstract

Biological neural circuits contain both unidirectional and reciprocal connections, but how reciprocity affects circuit computation remains unclear. We varied reciprocity \r{ho} across 21 levels in recurrent-network simulations while holding network size, connection density, and directed-edge count fixed. Low-reciprocity networks produced more diverse stimulus-driven activity and performed better on four continuous-input tasks, whereas higher reciprocity generally improved performance on three delayed-memory tasks. Increasing reciprocity reduced online memory capacity and accelerated recovery from perturbations. A post-cue noise experiment showed that higher reciprocity suppressed internal noise propagation but made cue-evoked states less distinguishable, and component matching separated the contributions of cue separation and within-cue dispersion to readout robustness. Within this ensemble, reciprocity thus mediates a three-way trade-off among representational capacity, local stability, and task-state separability: higher reciprocity lowers capacity, strengthens stability, and compresses separability. Whether it improves behaviour depends on whether the stability gain outweighs the lost separation between task states. Connectome and microcircuit datasets from five species showed intermediate reciprocity, while reciprocity varied substantially across adult Drosophila brain regions, providing compatible structural context and motivating circuit-specific functional tests.

Explore related subjects

q-bio.NC

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Xiaoyu Zhang, Mingyi Huang, Yixuan Cui, Xin Lu, Yuguo Yu. 2026-10-14. Reciprocity coordinates a three-way trade-off in recurrent neural circuits. https://arxiv.org/abs/2610.19949

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Neural Computation Runs on Dissipative Structures Under Cyclic Calibration: A Systems Perspective on Sleep

Biological function operates far from thermodynamic equilibrium, sustained by metastable structures that continuously cycle through synthesis, function execution, and degradation--consuming energy and exporting entropy in the process. Neural computation exemplifies this dissipative organization, relying on low-entropy structures engaged in dynamic processes that define key information transfer and processing parameters such as gain, threshold, and integration time. This perspective proposes that sleep serves as a systemic recalibration phase required to reset crucial signaling parameters that drift during dissipative processing. During wakefulness, neural activity progressively consumes and disrupts the organization of the very substrate from which it emerges, leading to parameter drift and reduced computational precision. Sleep maintains computational accuracy across activity cycles by recalibrating function-structure coupling--that is, by counterbalancing entropic drift. Rather than explaining the experimentally observed sleep need through homeostatic regulation of unknown physiological variables, this perspective reframes it as an emergent property of neural computation under thermodynamic constraints.

q-bio.NC↗
arXiv2026

Transferable conditional brain charts link cortico-subcortical structure to Alzheimer's disease progression

Normative models chart how brain measures vary across the lifespan, expressing each individual as a deviation from a healthy reference. However, most regional models of cortical thickness disregard dependencies between cortical and subcortical anatomy, and transferring them to a new site typically requires retraining or a recalibration of location and scale alone. We condition lifespan norms of cortical thickness on subcortical volume, such that each deviation indicates whether an individual's cortex lies within the healthy range given their subcortical anatomy, and make them transferable. In 21,351 healthy individuals from 216 sites, conditioning on ten subcortical structures improved out-of-sample prediction on average, most for the amygdala and hippocampus. Refitting only the intercepts of all four distributional parameters on about 20 healthy individuals, as implemented in our open-source package NormGAMLSS, calibrated most of six unseen sites without retraining, for base and conditional models. In the Alzheimer's Disease Neuroimaging Initiative, global deviation and structural dissociation (the reduction in mean absolute deviation under conditioning) predicted conversion to mild cognitive impairment and dementia beyond the Clinical Dementia Rating and dissociation added beyond global deviation for dementia, even at similar clinically recorded impairment. Conditional norms thus provide relevant, individual-level information.

q-bio.NC↗
arXiv2026

State Space Models Naturally Produce Time Cell and Oscillatory Behaviors and Scale to Abstract Cognitive Functions

A grand challenge in modern neuroscience is bridging the gap between the microscale neural circuits and cognitive functions. While extensive knowledge exists about neuronal connectivity and biophysics, how low-level phenomena give rise to abstract behaviors remains largely unresolved. Here, we propose that a mathematical framework based on State Space Models offers mathematical sufficiency and computational plausibility to generate response profiles that resemble time cells, ramping activity, and oscillatory responses. Our results suggest that such behaviors can emerge from a structured multi-timescale complex rotational basis, which enables optimized history compression and naturally generates structured temporal dynamics even prior to training, mirroring recent findings in biological circuits. Learning exploits and adapts to the pre-existing temporal basis for specific tasks. We show that the model can readily scale to abstract cognitive functions such as event counting, supporting its use as a computationally tractable framework for understanding neural activities.

q-bio.NC↗