prfmodel.fitters.grid_values_1d_prf

prfmodel.fitters.grid_values_1d_prf(stimulus: prfmodel.stimuli.PRFStimulus, num_mu: int = 49, mu_extent: float = 2.0, num_sigma: int = 50, sigma_range: tuple[float, float] | None = None, log_sigma: bool = True) → dict[str, numpy.ndarray]

Create grid search values for the center and size of a one-dimensional pRF from the stimulus.

The values for the pRF center mu span the stimulus coordinates, extended beyond them by mu_extent around their midpoint. The values for the pRF size sigma range from the smallest spacing to the full range of the stimulus coordinates by default.

Parameters:
  • stimulus (PRFStimulus) – One-dimensional stimulus. The coordinates do not need to be regularly spaced.

  • num_mu (int, optional) – Number of values for mu.

  • mu_extent (float, optional) – Range of the pRF center values relative to the range of the stimulus coordinates. With 1.0, the values span exactly the stimulus; with 2.0 (the default), they span twice its range around its midpoint.

  • num_sigma (int, optional) – Number of values for sigma.

  • sigma_range (tuple[float, float] or None, optional) – Smallest and largest value for sigma. If None, the smallest value is the smallest spacing between the stimulus coordinates and the largest value is their range.

  • log_sigma (bool, optional) – Whether the values for sigma are log-spaced (the default) or linearly spaced. Log spacing samples small pRFs, whose predictions change quickly with their size, more densely.

Returns:

Values for mu and sigma that can be passed to prfmodel.fitters.GridFitter.fit(). Values for the other model parameters (e.g., baseline and amplitude) must be added, for example, with the | operator.

Return type:

dict[str, numpy.ndarray]

Raises:

ValueError – If the stimulus is not 1-dimensional, num_mu is smaller than one, mu_extent is not positive, or the lower limit of sigma_range is not positive when log_sigma=True.

Warns:

GridAlignmentWarning – If the edges or the midpoint of the stimulus do not fall on the values for mu. The warning suggests a value for num_mu that aligns them.

See also

grid_values_2d_prf

Grid search values for two-dimensional pRFs.

Examples

>>> import numpy as np
>>> from prfmodel.examples import load_1d_prf_lognumerosity_stimulus
>>> stimulus = load_1d_prf_lognumerosity_stimulus()  # Log numerosities from log(1) to log(20)
>>> values = grid_values_1d_prf(stimulus, num_mu=9, num_sigma=5)
>>> np.round(values["mu"], 2)
array([-1.5 , -0.75,  0.  ,  0.75,  1.5 ,  2.25,  3.  ,  3.74,  4.49])
>>> np.round(values["sigma"], 2)
array([0.15, 0.32, 0.68, 1.43, 3.  ])