How to avoid event detection of initial conditions in scipy solve_ivp

I am working on a simulation of neural circuits. The circuit consists of eight segments, and I solve the activity of each segment with differential equations. The differential equation is

dE[i]/dt = f(b[i] * E[i]) (i = 0, 1, ... 7)

and b[i] is changed when E[i] cuts the threshold from below to above. To implement this, I built the following code using scipy's solve_ivp.

seg_number = 8
t0 = 0
tend = 2e5
x0 = [0]*seg_number

def gen_event(i):    
    def event(t, x):
        return x[i] - threshold
    event.terminal = True
    event.direction = 1
    return event

events = [gen_event(i) for i in range(seg_number)]


def func(t, x): #differential equation
    dx = []
    for i in range(seg_number):
    ・・・
    return(dx)


while True:
    orbit = solve_ivp(func, (t0, tend), x0, events = events)
    if orbit.status == 1:
        ・・・  # update b[i]

        t0 = orbit.t[-1]
        x0 = orbit.y[:, -1].copy()

    else:
        break

However, in this code, when b[i] and the initial conditions are changed and the differential equations are solved again, the initial condition is detected as an event, causing an infinite loop. If I add a small amount to the initial conditions, the solution proceeds, but I don't think this is a good solution.

x0 = orbit.y[:, -1].copy() + 1e-12