The actual paper is available [pdf]: http://dx.doi.org/10.1016/j.cub.2020.05.054

[Edit] Notes:

• People tested have all 3 cones but reduced function of one type (‘anomolous trichromats’) — either green (‘deuteranomoly’) or red (‘protoanomoly’). These are the most common forms of colour blindness.

• “All tests were performed without the glasses.” Hypothesis that the brain is being trained to compensate for the weak cones, like an audio equalizer [my analogy].

• They were testing a (patented) commercial product, in which one of the authors is invested. There may be specifics of the notch filters in the patent that were not in the paper. (I'm at work and not allowed to read active patents.)