Neighbour-sensing model¶
A virtual hyphal-growth model in which tip directions respond to fields from neighboring mycelium.
Core Idea¶
The Neighbour-Sensing model simulates fungal hyphal growth through virtual tips in a three-dimensional mycelial network. A tip's direction changes with abstract field information from surrounding modeled hyphae and other declared growth influences. As virtual segments extend, they change the neighbor context for later updates, allowing collective network or fruit-body patterns to emerge. This is not a static fungus drawing or a generic branching animation without neighbor-derived influence.
Meškauskas and colleagues' published 2004 work used the model to generate virtual fruit-body forms and to simulate colonies in semi-solid settings such as agar or soil. The colony study is an attested research application, not evidence that living fungi literally compute the stipulated mathematical field. The virtual network represents selected spatial growth relations while omitting much physiology. Image resemblance may motivate comparison with organisms, but species-specific mechanism or prediction requires independent validation. The exact model therefore sits under Representation yet remains a specialized fungal simulation.
Scope of Application¶
These uses retain virtual hyphal tips and neighbor-derived field updates.
- Fungal morphogenesis research. Explore how local tip rules can generate network and fruit-body patterns.
- In-silico colony comparison. Compare modeled colonies under declared substrate and tropism assumptions.
- Computational biology. Represent local interactions and resulting spatial structures without equating model with mechanism.
- Model evaluation. Ask which simulated morphological features match observations and which remain assumptions.
Clarity¶
Identify virtual mycelium, active tip states, neighbor-derived field values, and tip-direction updates. A generic branching animation is the nearest miss if it omits the neighbor field. The published colony simulations show what the virtual rules can generate, not proof that the abstract field is a literal fungal signal. Keep morphology output separate from biological mechanism.
Manages Complexity¶
The model compresses many hyphal interactions into field values and local vector updates. That permits exploration of collective morphology without tracking every physiological process. It also makes a limit visible: a generated network has fewer causal commitments than a living fungus. Separating geometric output from biological explanation lets a useful simulation remain useful without inflating its inferential status.
Abstract Reasoning¶
- Specify the virtual hyphal network and which tips remain active.
- State how neighboring modeled segments influence each tip's abstract field.
- Trace tip-direction/growth updates and the resulting spatial pattern.
- Compare changed rule scenarios without treating image similarity as a biological mechanism.
- Qualify any real-fungus inference by independent morphological or physiological evidence.
Knowledge Transfer¶
The neighbor-field/tip-update structure transfers among modeled fungal colonies or fruit-body simulations when the same virtual roles are retained and parameters declared. An agar-like in-silico colony does not automatically predict soil behavior or a particular species. Other agent-based growth simulations share a local-to-global skeleton, but if they lack this specific neighbor-derived hyphal field they are analogies, not instances. The broader target/medium/fidelity relation is a strict Representation instance.
Relationships to Other Abstractions¶
Current abstraction Neighbour-sensing model Domain-specific
Parents (1) — more general patterns this builds on
-
Neighbour-sensing model is a kind of Representation Prime
Virtual hyphal tips and neighbor fields map selected living-growth relations into a bounded simulation medium.
Hierarchy path (1) — routes to 1 parentless root
- Neighbour-sensing model → Representation → Abstraction
Neighborhood in Abstraction Space¶
Neighbour-sensing model sits in a sparse region of the domain-specific corpus (72nd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Cellular & Evolutionary Biological Processes (16 abstractions)
Nearest neighbors
- Network mapping — 0.84
- Graphical Models for Protein Structure — 0.84
- Algebraic Surface — 0.83
- Graph dynamical system — 0.83
- Prim’s Algorithm — 0.83
Computed from structural-signature embeddings · 2026-10-08