Skip to main content
. 2014 Jan 19;369(1634):20120397. doi: 10.1098/rstb.2012.0397

Figure 2.

Figure 2.

Simulations of learning to read through phonological decoding and self-teaching. Performance of the network using five different word recognition thresholds (0.05, 0.15, 0.25, 0.35 and 0.45). (a) Percentage of learnt/lexicalized words; (b) numbers of neighbours in the cohort of each recognized word and (c) percentage of items in each cohort where the correct item was also the most strongly activated item.