Source distance from centre

Describe the new feature or enhancement

Following PR #8534 ("source depth"), I would find it useful to have a method for source spaces compute_distance_to_center() which returns each vertex' distance to an origin of vertex coordinates. This could be useful to evaluate source/head/forward models without specifying a particular sensor configuration.
A possible API could be:

def compute_distance_to_centre(src, origin=None):
    """Compute distances between vertices and their origin.

    Parameters
    ----------
    src : instance of SourceSpaces
        The object with vertex positions for which to compute distances to origin.
    origin: None | str
        The origin of vertex coordinates to use for distance computations. Can be
        * None (default): Compute Euclidean norm of vertex coordinates (i.e. origin [0, 0, 0].
        * Talairach: Compute Euclidean norm of vertex coordinates in Talairach coordinate system.
        * MNI: Compute Euclidean norm of vertex coordinates in Talairach coordinate system.
        * COG: Compute Euclidean norm of vertex coordinates with respect to their center of gravity.

    Returns
    -------
    distances : array of shape (n_vertices,)
        The distances (in metres) from origin to vertex locations.
    """

Describe your proposed implementation

Ideally as method for source space objects.

Describe possible alternatives

Could also be a function in source_space.py (similar to compute_distance_to_sensors).

Additional comments

The different options would require transformations to MNI and Talairach coordinate systems. I would need help with that.

Start in source_space.py and review compute_distance_to_sensors and PR #8534 on source depth. Clarify the source-space API and the supported origins, especially Talairach, MNI, and COG transformations. Done means source spaces can return per-vertex distances in metres with documented behavior and coverage for the supported origin options.