Meeting Abstract

P2.78  Monday, Jan. 5  Linear Modeling of Tracking Behavior in Weakly Electric Fish ZHUANG, K.*; ROTH, E.; FORTUNE, E.S.; COWAN, N.J.; Johns Hopkins University; Johns Hopkins University; Johns Hopkins University; Johns Hopkins University kzhuang1@jhu.edu

Weakly electric knifefish swim to maintain their relative position within a moving refuge. Dynamical system models provide a language for describing such systems. When applicable, linear dynamical systems models admit a particularly powerful suite of analyses that facilitates the generation of novel hypotheses. Here, we present a methodology for assessing the regimes for which linear dynamical systems approximate experimentally measured sensorimotor performance. Using this methodology, we tested the linearity hypothesis for refuge tracking in several species of weakly electric knifefish (Eigenmannia virescens, Apteronotus leptorhynchus, Apteronotus albifrons) for refuge movements limited to approximately 2 Hz oscillation frequency and 15 cm/s maximum velocity. We measured the system output (fish position and velocity) to an array of input signals (refuge trajectories), including pure sinusoids, sums of sinusoids, and band-limited noise. Tools for data analysis included ARMAX (autoregressive moving average with exogenous input) modeling and coherence analysis. ARMAX modeling provided a way to statistically quantify the linear dynamics of tracking responses using model parameters. Coherence analysis, commonly used for characterizing neural activity, provided a way to assess frequency regimes in which tracking behavior was linear. The results were consistent with the linearity hypothesis proposed in Cowan and Fortune (2007).