Computational modeling reveals that the linear and nonlinear modulation of action selection by direct versus indirect pathway qualitatively hold with additional striatal collateral connections.
(A) Schematic for ‘Triple-control’ model with D1-D1 collateral connections. (B) Correct rate change in 2s trials (upper panel) and 8s trials (bottom panel) when manipulating D1-SPNs with different manipulation strengths (n = 10, one-way repeated-measures ANOVA, effect of manipulation strength, 2s trials: F40,369 = 1.328, p = 0.0945; 8s trials: F40,369 = 7.595, p < 0.0001). Green lines: ‘Triple-control’ model with D1-D1 collateral connections. Gray lines: the same simulation results as shown in Figure 7(C, G). (C) Correct rate change in 2s trials (upper panel) and 8s trials (bottom panel) when manipulating D2-SPNs with different manipulation strengths (n = 10, one-way repeated-measures ANOVA, effect of manipulation strength, 2s trials: F40,369 = 38.22, p < 0.0001; 8s trials: F40,369 = 34.29, p < 0.0001). Green lines: ‘Triple-control’ model with D1-D1 collateral connections. Gray lines: the same simulation results as shown in Figure 7(D, H). (D) Schematic for ‘Triple-control’ model with D1-D2 collateral connections. (E) Correct rate change in 2s trials (upper panel) and 8s trials (bottom panel) when manipulating D1-SPNs with different manipulation strengths (n = 10, one-way repeated-measures ANOVA, effect of manipulation strength, 2s trials: F40,369 = 0.9335, p = 0.5893; 8s trials: F40,369 = 8.778, p < 0.0001). Purple lines: ‘Triple-control’ model with D1-D2 collateral connections. Gray lines: the same simulation results as shown in Figure 7(C, G). (F) Correct rate change in 2s trials (upper panel) and 8s trials (bottom panel) when manipulating D2-SPNs with different manipulation strengths (n = 10, one-way repeated-measures ANOVA, effect of manipulation strength, 2s trials: F40,369 = 40.94, p < 0.0001; 8s trials: F40,369 = 26.61, p < 0.0001). Purple lines: ‘Triple-control’ model with D1-D2 collateral connections. Gray lines: the same simulation results as shown in Figure 7(D, H). (G) Schematic for ‘Triple-control’ model with D2-D1 collateral connections. (H) Correct rate change in 2s trials (upper panel) and 8s trials (bottom panel) when manipulating D1-SPNs with different manipulation strengths (n = 10, one-way repeated-measures ANOVA, effect of manipulation strength, 2s trials: F40,369 = 0.6827, p = 0.9299; 8s trials: F40,369 = 10.06, p < 0.0001). Blue lines: ‘Triple-control’ model with D1-D2 collateral connections. Gray lines: the same simulation results as shown in Figure 7(C, G). (I) Correct rate change in 2s trials (upper panel) and 8s trials (bottom panel) when manipulating D2-SPNs with different manipulation strengths (n = 10, one-way repeated-measures ANOVA, effect of manipulation strength, 2s trials: F40,369 = 153.3, p < 0.0001; 8s trials: F40,369 = 38.38, p < 0.0001). Blue lines: ‘Triple-control’ model with D1-D2 collateral connections. Gray lines: the same simulation results as shown in Figure 7(D, H).