Figure 3

Development and maturation of fitted model parameters. (a) Dynamic model of binocular rivalry, with competition, adaptation, and noise. Two representations (\(r_1\), \(r_2\)) are driven by associated visual inputs (\(I_1\), \(I_2\)) and independent noise (\(n_1\), \(n_2\)). Each representation is inhibited by the other, as well as by the associated adaptive state (\(a_1\), \(a_2\)). Free model parameters are competition strength (\(\beta\)), adaptation strength (\(\phi _a\)), adaptation time-scale (\(\tau _a\)), and noise amplitude (\(\sigma _n\)). Input strength is fixed at \(I_1=I_2=1\). (b) Representative example of model dynamics with abrupt dominance reversals of activities \(r_1\) and \(r_2\), gradual build-up and recovery of adaptive states \(a_1\) and \(a_2\) (middle), and noise \(n_1\) and \(n_2\) added to visual input \(I=1\) (\(\beta =2\), \(\phi _a=0.7\), \(\tau _a=0.3\), \(\sigma _n=0.2\)). Reversals may be triggered by differential adaptive state \(a_1-a_2\), differential noise \(n_1-n_2\), or both. (c, d) Parameter triplets \(\phi _a\)-\(\tau _a\)-\(\sigma _n\) fitted to reproduce (within a ±5%) experimentally observed mean distribution parameters. Competition strength was fixed at \(\beta =3\). Centroid values (coloured dots) and confidence range (± standard deviation, grey mesh). Red stars represent peak age of maturation index (Fig. 2d).