Constraints for initial conditions in link function

[RaoXie]

Dear Max and pyddm user,

I am modelling behavioral data using EEG data that spans a range from negative infinity to infinity, for certain parameters, such as, bound (must be positive), starting point (value range [-1, 1]?), I am wondering how to appropriately constrain the parameters in the model? Should I:
Apply constraints to the link function to ensure values stay within valid ranges?
Or, map the output of the link function to another function that keeps the results within an acceptable range? If yes, are there any recommendations or empirically more favorable transformation functions for them?

E.g., I tried to fit starting point that varies according to the EEG_signal, but of course it failed because of the initial conditions:

def x0_link(EEG_signal, beta1, beta0):
        x0 = beta1 * EEG_signal + beta0
        return x0
    m = pyddm.gddm(starting_position=x0_link,
                drift=v_link, 
                noise="noise", 
                nondecision="nd",
                parameters={
                    "beta0": (-1, 1),
                    "beta1": (-5, 5),
                    "noise": (.1, 2),
                    "nd": (0, 2)
                },
                conditions=["EEG_signal"],
...
                )

Thank you so much for your help and time again!

Best, Rao