Table 2. Use of Internal Computational Parameters to Predict Developmental Outcomes (Persisting Delay, Resolving Delay) in a Polygenic Model of Language Delay.
| Computational parameter | Processing role | Effect size of PD versus RD comparison |
|---|---|---|
| Note. PD = persisting delay; RD = resolving delay. Scores show ηp2 effect sizes from ANOVA comparing PD and RD groups (see Thomas & Knowland, 2014, Table 2, for parallel analyses using logistic regression methods). | ||
| * Effect reliable at p < .05. ** Effect reliable at p < .01. | ||
| Number of internal units | Capacity | .031** |
| Pruning threshold | Capacity/Regressive events | .021* |
| Learning algorithm | Capacity/Plasticity | .104** |
| Lexical-semantic learning rate | Plasticity | .024** |
| Unit discriminability | Plasticity/Signal | .025** |
| Processing noise | Signal | .026** |