nstat.simulate_cif_from_stimulus

nstat.simulate_cif_from_stimulus(time, stimulus, beta0, beta1, rng=None)[source]

Simulate a spike train from a log-linear CIF driven by a stimulus.

Computes the conditional intensity lambda(t) = exp(beta0 + beta1 * x(t)) in spikes/second, then draws spikes by independent per-bin Bernoulli sampling (p = 1 - exp(-rate*dt)) at the time-grid resolution implied by time.

Parameters:

  • time (ndarray, shape (n_time,)) – Time vector in seconds.

  • stimulus (ndarray, shape (n_time,)) – Stimulus values at each time sample.

  • beta0 (float) – Log-baseline rate (intercept of the log-linear CIF).

  • beta1 (float) – Stimulus coefficient.

  • rng (np.random.Generator, optional) – Random generator. Defaults to np.random.default_rng().

Returns:

  • spike_train (nspikeTrain) – Simulated spike train.

  • rate_hz (ndarray, shape (n_time,)) – The instantaneous CIF in spikes per second.

  • log_rate (ndarray, shape (n_time,)) – The log-rate beta0 + beta1 * x(t) (useful for diagnostics).

Return type:

tuple[nspikeTrain, ndarray, ndarray]