Fig. 1. Spiketrains and PRC of the ML model when I = 45.5 µA·cm^–2, A = 1.65 µA·cm^–2, and d = 4.4 ms: (a) Square pulse disturbance current (dashed line), and spike trains without (dotted line) and with square pulse disturbance (solid line); (b) PRC. 当I = 45.5 µA·cm^–2, 方波脉冲幅值A = 1.65 µA·cm^–2, 宽度d = 4.4 ms时, ML模型的放电序列和PRC　(a) 方波脉冲刺激电流(短划线)、没有方波脉冲的放电(点线)和有方波脉冲的放电(实线); (b) PRC

Fig. 2. (a) Bifurcation of ML model with respect to I; (b) spike trains of ML model when I = 45.5 µA·cm^–2; (c) resting state of ML model when I = 44 µA·cm^–2; (d) the changes of ISIs (solid line) and frequency (dashed line) with respect to I. (a) ML模型随I的平衡点分岔; (b) 当I = 45.5 µA·cm^–2时, ML模型的放电序列图; (c) 当I = 44 µA·cm^–2时, ML模型处于静息状态; (d) ISIs和频率随I的变化(实线表示ISIs, 虚线表示频率)

Fig. 3. Spike trains induced by square pulse current applied at different phases when I = 45.5 µA·cm^–2. The spike trains (solid line) influenced by negative square pulse current (dashed line) and the trains (red dotted line) without negative square pulse current. (a) t_s = 22 ms, A = –0.6 µA·cm ^–2, d = 4.9 ms; (b) t_s = 22 ms, A = –1.65 µA·cm ^–2, d = 4.8 ms; (c) t_s = 40 ms, A = –0.6 µA·cm ^–2, d = 4.9 ms; (d) t_s = 40 ms, A = –1.65 µA·cm ^–2, d = 4.8 ms. 当I = 45.5 µA·cm^–2时, 负向方波脉冲电流(虚线)作用在不同相位的放电序列(实线)和无方波脉冲作用的放电序列(红色的点线)　(a) t_s = 22 ms, A = –0.6 µA·cm ^–2, d = 4.9 ms; (b) t_s = 22 ms, A = –1.65 µA·cm ^–2, d = 4.8 ms; (c) t_s = 40 ms, A = –0.6 µA·cm ^–2, d = 4.9 ms; (d) t_s = 40 ms, A = –1.65 µA·cm ^–2, d = 4.8 ms

Fig. 4. PRC induced by negative square pulse current near the Hopf bifurcation point in the ML model when I = 45.5 µA·cm^–2: (a) A = –0.6 µA·cm ^–2, d = 4.9 ms; (b) A = –1.65 µA·cm ^–2, d = 4.8 ms. 当I = 45.5 µA·cm^–2时, ML模型在Hopf分岔点附近的负向脉冲刺激诱发的PRC　(a) A = –0.6 µA·cm ^–2, d = 4.9 ms; (b) A = –1.65 µA·cm ^–2, d = 4.8 ms

Fig. 5. Inhibitory autapse current (dashed line) and spike trains (solid line) of the ML model with inhibitory autapse when I = 45.5 µA·cm^–2 and g_aut = 0.04 mS·cm^–2: (a) τ = 0 mS; (b) τ = 10 mS; (c) τ = 20 mS; (d) τ = 30 mS; (e) τ = 40 mS; (f) τ =50 mS. 当I = 45.5 µA·cm^–2, g_aut = 0.04 mS·cm^–2, 具有抑制性自突触ML模型的放电模式(实线)与抑制性自突触电流(短划线) (a) τ = 0 mS; (b) τ = 10 mS; (c) τ =20 mS; (d) τ = 30 mS; (e) τ = 40 mS; (f) τ =50 mS

Fig. 6. Change of normalized ISIs(boldsolid line) and firing frequency (thin solid line) with respect to time delay τ: (a) g_aut = 0.01 mS·cm^–2; (b) g_aut = 0.04 mS·cm^–2. 不同耦合强度g_aut下归一化的ISIs (粗实线)和放电频率(细实线)随时滞τ的变化　(a) g_aut = 0.01 mS·cm^–2; (b) g_aut = 0.04 mS·cm^–2

Fig. 7. Dependence of average ISIs on time delay τ and coupling strength g_aut when D = 0.5 µA·cm^–2. 当噪声强度D = 0.5 µA·cm^–2时, 平均ISIs对时滞τ和耦合强度g_aut的依赖关系

Fig. 8. (a) Dependence of standard deviation of ISIs(STD) on time delay τ and coupling strength g_aut; (b) the dependence of coefficient of variation of ISIs(CV) on delay τ and coupling strength g_aut. The parameter values are I = 45.5 µA·cm^–2 and D = 0.5 µA·cm^–2. 当I = 45.5 µA·cm^–2, D = 0.5 µA·cm^–2时, (a) ISIs的STD对时滞τ和耦合强度g_aut的依赖关系, (b) ISIs的CV对时滞τ和耦合强度g_aut的依赖关系

Fig. 9. Changes of coefficient of variation (CV) of ISIs with respect to coupling strength g_aut when time delay τ is fixed at different values: (a) τ = 27 ms; (b) τ = 30 ms; (c) τ = 40 ms; (d) τ = 46 ms. 固定时滞τ在不同水平下, ISIs的CV随着耦合强度g_aut的变化　(a) τ = 27 ms; (b) τ = 30 ms; (c) τ = 40 ms; (d) τ = 46 ms

Fig. 10. Changes of coefficient of variation (CV) of ISIs with respect to time delay τ when coupling strength g_aut is fixed at different levels: (a) g_aut = 0.31 mS·cm^–2; (b) g_aut = 0.61 mS·cm^–2. 固定耦合强度g_aut在不同水平下, ISIs的变异系数CV随着时滞τ的变化　(a) g_aut = 0.31 mS·cm^–2; (b) g_aut = 0.61 mS·cm^–2

Fig. 11. Effect of time delay τ on spike-timing precision of neuron model when I = 45.5 µA·cm^–2, D = 0.5 µA·cm^–2, and g_aut = 0.61 mS·cm^–2: (a) τ = 1 ms; (b) τ = 10 ms; (c) τ = 30 ms; (d) τ = 50 ms. 当I = 45.5 µA·cm^–2, D = 0.5 µA·cm^–2, 耦合强度g_aut = 0.61 mS·cm^–2时, 时滞τ对神经元模型的精确放电的影响　(a) τ = 1 ms; (b) τ = 10 ms; (c) τ = 30 ms; (d) τ = 50 ms