. 2016 Oct 18;10:181. doi: 10.3389/fnbeh.2016.00181

Table 3.

Simulations with two manipulanda: test phase.

	No food devalued				Second food devalued
	Mean actions		t-test		Mean actions		t-test
	L1	L2			L1	L2
CONTROL	7.96	10.43	t₍₃₉₎ = −1.59	p = 0.116	25.11	0.25	t₍₃₉₎ = 91.22	p < 0.001
BLA/IC-pre	9.43	8.91	t₍₃₉₎ = 0.39	p = 0.695	9.34	8.71	t₍₃₉₎ = 0.46	p = 0.645
NAcCO-pre	8.30	9.73	t₍₃₉₎ = −1.16	p = 0.252	8.49	9.54	t₍₃₉₎ = −0.85	p = 0.398
DMS-pre	8.81	8.60	t₍₃₉₎ = 0.28	p = 0.777	8.69	8.88	t₍₃₉₎ = −0.30	p = 0.765
PL-pre	9.38	8.80	t₍₃₉₎ = 0.42	p = 0.678	8.95	8.91	t₍₃₉₎ = 0.03	p = 0.978
BLA-post	9.13	9.54	t₍₃₉₎ = −0.28	p = 0.782	9.25	9.56	t₍₃₉₎ = −0.22	p = 0.829
NAcCO-post	9.18	8.58	t₍₃₉₎ = 0.41	p = 0.684	9.08	8.80	t₍₃₉₎ = 0.18	p = 0.855
DMS-post	9.56	7.54	t₍₃₉₎ = 1.96	p = 0.054	9.16	8.08	t₍₃₉₎ = 1.08	p = 0.285
PL-post	9.61	8.56	t₍₃₉₎ = 0.59	p = 0.554	25.09	0.35	t₍₃₉₎ = 75.22	p < 0.001

For each condition the means of actions on the first lever (L1) and second lever (L2), averaged over 40 replications of the simulation, are reported. The left columns refer to tests without devaluation. The right columns refer to tests where the food associated to L2 had been devalued. The CONTROL row refers to simulations with the intact model, whereas the other rows refer to simulations where various structures of the model were lesioned either before training (-pre) or after training (-post).