bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown
Опубликована: Сен. 20, 2024
A bstract Background The continuous attractor network (CAN) model has been effective in explaining grid-patterned firing the rodent medial entorhinal cortex, with strong lines of experimental evidence and widespread utilities understanding spatial navigation path integration. surprising lacuna CAN analyses is paucity quantitative studies on impact afferent sensory noise Here, we evaluate accuracy position estimates derived from pattern flow velocity. Motivated by ability border cells to act as an error-correction mechanism, also assess interaction between cell inputs performance. Methodology We used established 2D that received velocity a virtual animal traversing arena generate firing. estimated activity compute activity-based estimate at each time step. tracked difference real positions function called it deviation integrated (DIP). defined be additive Gaussian, different levels achieved changing variance. introduced north east connected them grid based co-activity patterns. For noise, computed DIP metrics for presence vs . absence cells. Importantly, avoid potential bias owing use single trajectory computing these measurements, performed all simulations across 50 trajectories. Results scores (as DIP) showed pronounced trajectory-to-trajectory variability, even noise-free network. With introduction variability prevailed unveiled dichotomous activity. Specifically, low improved estimation without altering In contrast, high impaired well activity, although were more sensitive compared stochastic resonance observed relationship level was partially explained noisy rectification nonlinearity neural transfer function. Finally, levels, grid-score addition inputs. Across population trajectories, yielded modest changes both measurements levels. Implications Our demonstrate robustness models does not extend other functions model. Stochastic reference implies biological CANs could evolve yield optimal performance (path integration) systems. An important methodological implication emerges our observations critical need account Given nature conclusions are bound erroneous thereby warranting multiple Together, unveil roles improving obtained models.
Язык: Английский