Ant navigation: understanding the resilience and self-developing nature of mini-brains in interaction with their environment.

Introduction

Organisms differ from machines in their resilience: their capacity to spontaneously recover from defects. This is because organisms, contrary to machines, are self-developing systems that change over time, like ‘a melody that sings itself’.

Importance of Understanding Resilience

The plastic mechanisms that enable animal and human self-development and resilience are barely known, but understanding them would profoundly impact how we apprehend brains, behaviours, and their evolution.

Project Overview

RESILI-ANT will tackle this question using the stunning ability of solitary foraging ants to visually navigate in complex environments. Ant navigational behaviours develop through stages with strong learning components and can recover from sensory-motor alterations that would disable any machine.

Advantages of Ants as a Model

The advantage is that the full ontogeny of these behaviours unfolds in a couple of days, can be easily manipulated, and involves a brain that is numerically much simpler and better-known than any vertebrate.

Methodology

We will combine an ecological approach, complexity science, and modelling with state-of-the-art technologies to dissect the mechanisms underlying these behavioural developmental processes. Our methodology includes:

1. Tracking ant ontogeny under different scenarios.
2. Performing straightforward manipulations to disclose the underlying rules.
3. Using virtual reality to probe their ability to compensate for severe sensory-motor defects.

Neural Implementation

In parallel, we will explore how these processes are implemented – in light of our ever-increasing knowledge of insect circuits – by augmenting our current neural models with network plasticity and recurrent connections between brain areas.

Evolutionary Investigation

The evolutionary causes and consequences of such plastic neural topologies will be investigated using neuro-evolution algorithms selecting for self-developing agents that forage in reconstructed environments.

Conclusion

The dialogue between simulations and observation will move us towards a concrete understanding of the self-developing nature and their fundamental and societal impact.