Training & Analysis:
Ringo et al. (1994) and others take this as evidence of less interhemispheric communication, requiring "multiple passes" to overcome.
We know there will be an onset difference of 9 time-steps. What remains after we account for that?shifted original
Same as Ringo model, but:
Run three experiments:
|Noise reduces over training, roughly simulating axonal maturation.||
Over learning, the difference in lesion-induced error between noise and no-noise conditions decreases, showing that an initial bias for local (intrahemispheric) circuits disappears.