State-space model results.

<p>A) Whisker plots showing the change in the fitted state memory term (A) across training days. Data presented as inverse parameter values to reflect time to change state (in trials). Same plotting conventions as <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0047336#pone-0047336-g002" target="_blank">Figure 2a,b</a>. B) Heat map of average response prediction term (B) across training. C,D) Whisker plots of the error correction term (F) and state noise term (Q, in variance units) respectively. Same plotting conventions as <i>A</i>. E) Heat map of the average response bias term (D) from the output equation (Methods, Eq. 2). F) Whisker plot of the motor noise term (R, in variance units) across training day. Same plotting conventions as <i>A</i>.