We propose a framework based on the concept of “state space” or “phase space” in dynamical systems, which converts variables changing in time to a trajectory, e.g. 4 different “behaviors” of a pendulum become 4 curves.pic.twitter.com/LO9AiNPdi1
-
-
Prikaži ovu nit
-
Applying our framework to C. elegans behavior, we develop a method to reconstruct a state space from videos of freely moving worms. Our method is based on the idea that good state space reconstructions should allow us to predict the future for as long as possible.pic.twitter.com/oLr2hT0nxU
Prikaži ovu nit -
The low-dimensional reconstructed state space allows us to understand a great deal about worm movement, without knowing the equations of motion! It is globally structured into trajectory bundles corresponding to the canonical behaviors of forward, backward and turning.pic.twitter.com/HuRH85I6WA
Prikaži ovu nit -
At small scales though, the trajectories look messy, and long-term prediction of worm behavior is not possible. A pair of nearby points in state space start close, but go apart after some time. The state space allows us to systematically study behavioral variability in C. eleganspic.twitter.com/sTzwLDhhg1
Prikaži ovu nit -
Analysis of the geometry and topology of these trajectories reveals several hints of deterministic variability coming from chaotic dynamics. These include unstable periodic orbits, exponential divergence of initial conditions and most surprisingly, a symmetric Lyapunov spectrum.
Prikaži ovu nit -
Dynamical systems are well-suited to understand the complexity of C. elegans behavior, characterized by its long term unpredictability, multi scale structure, and paradoxical coexistence of stereotypy & variability. And we believe this is just the start.
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.