A virtuous cycle between invertebrate and robotics research: perspective on a decade of Living Machines research DOI Creative Commons
Michael Mangan, Dario Floreano, Kotaro Yasui

et al.

Bioinspiration & Biomimetics, Journal Year: 2023, Volume and Issue: 18(3), P. 035005 - 035005

Published: March 7, 2023

Many invertebrates are ideal model systems on which to base robot design principles due their success in solving seemingly complex tasks across domains while possessing smaller nervous than vertebrates. Three areas particularly relevant for designers: Research flying and crawling has inspired new materials geometries from bodies (their morphologies) can be constructed, enabling a generation of softer, smaller, lighter robots. walking insects informed the controlling motion control) adapting environment without costly computational methods. And research combining wet neuroscience with robotic validation methods revealed structure function core circuits insect brain responsible navigation swarming capabilities mental faculties) displayed by foraging insects. The last decade seen significant progress application extracted invertebrates, as well biomimetic robots better understand how animals function. This Perspectives paper past 10 years Living Machines conference outlines some most exciting recent advances each these fields before outlining lessons gleaned outlook next invertebrate research.

Language: Английский

Neurons from pre-motor areas to the Mushroom bodies can orchestrate latent visual learning in navigating insects DOI Creative Commons
Antoine Wystrach

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2023, Volume and Issue: unknown

Published: March 10, 2023

ABSTRACT Spatial learning is peculiar. It can occur continuously and stimuli of the world need to be encoded according some spatial organisation. Recent evidence showed that insects categorise visual memories as whether their gaze facing left vs. right from goal, but how such categorisation achieved during remains unknown. Here we analysed movements ants exploring around nest, used a biologically constrained neural model show parallel, lateralized acquired straightforwardly agent explore world. During learning, ‘left’ ‘right’ formed in different comportments (of mushroom bodies lobes) through existing lateralised dopaminergic feedback pre-motor areas (the lateral accessory receiving output path integration (in central complex). As result, organises ‘internally’, without expressed behaviour; therefore, views learnt (without suffering memory overload) while insect free randomly or using any other navigational mechanism. After this circuit produces robust homing performance 3D reconstructed natural habitat despite noisy recognition performance. Overall illustrates continuous bidirectional relationships between centres orchestrate latent produce efficient navigation behaviour.

Language: Английский

Citations

18
Unpacking the navigation toolbox: insights from comparative cognition DOI Creative Commons
Kathryn J. Jeffery, Ken Cheng, Nora S. Newcombe

et al.

Proceedings of the Royal Society B Biological Sciences, Journal Year: 2024, Volume and Issue: 291(2016)

Published: Feb. 7, 2024

The study of navigation is informed by ethological data from many species, laboratory investigation at behavioural and neurobiological levels, computational modelling. However, the are often species-specific, making it challenging to develop general models how biology supports behaviour. Wiener et al . outlined a framework for organizing results across taxa, called ‘navigation toolbox’ (Wiener al. In Animal thinking: contemporary issues in comparative cognition (eds R Menzel, J Fischer), pp. 51–76). This proposes that spatial hierarchical process which sensory inputs lowest level successively combined into ever-more complex representations, culminating metric or quasi-metric internal model world (cognitive map). Some animals, notably humans, also use symbolic representations produce an external representation, such as verbal description, signpost map allows communication information instructions between individuals. Recently, new discoveries have extended our understanding constructed, highlighting relationships bidirectional, with higher levels feeding back influence lower levels. light these developments, we revisit toolbox, elaborate incorporate findings. toolbox provides common within different taxa can be described compared, yielding more detailed, mechanistic generalized navigation.

Language: Английский

Citations

8
Neural circuits for goal-directed navigation across species DOI Creative Commons
Jayeeta Basu, Katherine I. Nagel

Trends in Neurosciences, Journal Year: 2024, Volume and Issue: 47(11), P. 904 - 917

Published: Oct. 10, 2024

Language: Английский

Citations

7
The neuronal building blocks of the navigational toolkit in the central complex of insects DOI
Keram Pfeiffer

Current Opinion in Insect Science, Journal Year: 2022, Volume and Issue: 55, P. 100972 - 100972

Published: Sept. 17, 2022

Language: Английский

Citations

28
CATER: Combined Animal Tracking & Environment Reconstruction DOI Creative Commons
Lars Haalck, Michael Mangan, Antoine Wystrach

et al.

Science Advances, Journal Year: 2023, Volume and Issue: 9(16)

Published: April 21, 2023

Quantifying the behavior of small animals traversing long distances in complex environments is one most difficult tracking scenarios for computer vision. Tiny and low-contrast foreground objects have to be localized cluttered dynamic scenes as well trajectories compensated camera motion drift multiple lengthy recordings. We introduce CATER, a novel methodology combining an unsupervised probabilistic detection mechanism with globally optimized environment reconstruction pipeline enabling precision behavioral quantification natural environments. Implemented easy use highly parallelized tool, we show its application recover fine-scale trajectories, registered high-resolution image mosaic reconstruction, naturally foraging desert ants from unconstrained field By bridging gap between laboratory experiments, gain previously unknown insights into ant navigation respect motivational states, previous experience, current provide appearance-agnostic method applicable study wide range terrestrial species under realistic conditions.

Language: Английский

Citations

14
Neural Networks for Navigation: From Connections to Computations DOI Creative Commons
Rachel I. Wilson

Annual Review of Neuroscience, Journal Year: 2023, Volume and Issue: 46(1), P. 403 - 423

Published: July 10, 2023

Many animals can navigate toward a goal they cannot see based on an internal representation of that in the brain's spatial maps. These maps are organized around networks with stable fixed-point dynamics (attractors), anchored to landmarks, and reciprocally connected motor control. This review summarizes recent progress understanding these networks, focusing studies arthropods. One factor driving is availability Drosophila connectome; however, it increasingly clear navigation depends ongoing synaptic plasticity networks. Functional synapses appear be continually reselected from set anatomical potential interaction Hebbian learning rules, sensory feedback, attractor dynamics, neuromodulation. explain how space rapidly updated; may also brain initialize goals as fixed points for navigation.

Language: Английский

Citations

13
A model of cue integration as vector summation in the insect brain DOI Creative Commons
Robert Mitchell, Shahrzad Shaverdian, Marie Dacke

et al.

Proceedings of the Royal Society B Biological Sciences, Journal Year: 2023, Volume and Issue: 290(2001)

Published: June 26, 2023

Ball-rolling dung beetles are known to integrate multiple cues in order facilitate their straight-line orientation behaviour. Recent work has suggested that integrated according a vector sum, is, compass represented by vectors and summed give combined estimate. Further, cue weight (vector magnitude) appears be set reliability. This is consistent with the popular Bayesian view of integration: reduce or minimize an agent's uncertainty about external world. Integration believed occur at input insect central complex. Here, we demonstrate model head direction circuit complex, including plasticity synapses, can act as substrate for integration summation. show influence not necessarily driven Finally, present beetle behavioural experiment which, combination simulation, strongly suggests these do We suggest alternative strategy whereby weighted relative contrast, which also explain previous results.

Language: Английский

Citations

12
Neural circuit mechanisms for steering control in walking Drosophila DOI Open Access

Aleksandr Rayshubskiy,

Stephen L. Holtz,

Alexander Shakeel Bates

et al.

Published: May 2, 2025

Orienting behaviors provide a continuous stream of information about an organism’s sensory experiences and plans. Thus, to study the links between sensation action, it is useful identify neurons in brain that control orienting behaviors. Here we describe descending Drosophila predict influence orientation (heading) during walking. We show these cells have specialized functions: whereas one cell type predicts sustained low-gain steering, other transient high-gain steering. These latter integrate internally-directed steering signals from head direction system with stimulus-directed multimodal pathways. The inputs are organized produce “see-saw” commands, so increasing output hemisphere accompanied by decreasing hemisphere. Together, our results internal external drives integrated motor commands different timescales, for flexible precise space.

Language: Английский

Citations

0
How the insect central complex could coordinate multimodal navigation DOI Creative Commons
Xuelong Sun, Shigang Yue, Michael Mangan

et al.

eLife, Journal Year: 2021, Volume and Issue: 10

Published: Dec. 9, 2021

The central complex of the insect midbrain is thought to coordinate guidance strategies. Computational models can account for specific behaviours, but their applicability across sensory and task domains remains untested. Here, we assess capacity our previous model (Sun et al. 2020) visual navigation generalise olfactory its coordination with other in flies ants. We show that fundamental this use a biologically plausible neural copy-and-shift mechanism ensures information presented format compatible steering circuit regardless source. Moreover, same shown allow transfer cues from unstable/egocentric stable/geocentric frames reference, providing first by which foraging insects robustly recover environmental disturbances. propose these circuits be flexibly repurposed different navigators address unique ecological needs.

Language: Английский

Citations

25
An insect-inspired model facilitating autonomous navigation by incorporating goal approaching and collision avoidance DOI Creative Commons
Xuelong Sun, Qinbing Fu, Jigen Peng

et al.

Neural Networks, Journal Year: 2023, Volume and Issue: 165, P. 106 - 118

Published: May 24, 2023

Being one of the most fundamental and crucial capacity robots animals, autonomous navigation that consists goal approaching collision avoidance enables completion various tasks while traversing different environments. In light impressive navigational abilities insects despite their tiny brains compared to mammals, idea seeking solutions from for two key problems navigation, i.e., avoidance, has fascinated researchers engineers many years. However, previous bio-inspired studies have focused on merely these at time. Insect-inspired algorithms synthetically incorporate both investigate interactions mechanisms in context sensory–motor closed-loop are lacking. To fill this gap, we propose an insect-inspired algorithm integrate mechanism as global working memory inspired by sweat bee's path integration (PI) mechanism, model local immediate cue built upon locust's lobula giant movement detector (LGMD) model. The presented is utilized drive agents complete task a manner within bounded static or dynamic environment. Simulation results demonstrate synthetic capable guiding agent challenging robust efficient way. This study takes first tentative step insect-like with functionalities (i.e., interrupt) into coordinated control system future research avenues could build upon.

Language: Английский

Citations

10