With little daylight, artificial light does most of the work, and whatever colour that light has becomes the room's colour. Everything also reads darker and less distinct: the eye's ability to separate similar colours falls as light levels drop, so two shades that are clearly different on a well-lit sample board can become indistinguishable on the wall. Saturation suffers most, because a saturated colour needs light to return its colour at all.
Pale, warm-to-neutral colours with lower visual weight are the conventional choice, and they do genuinely reflect more of the little light available. But the opposite approach is equally valid and often more successful: rather than fighting a lack of light with pale colour that ends up looking grey, commit to a deep, warm colour and light the room properly with lamps. A small dark room can be a good dark room; it rarely becomes a bright one.
Deep, highly saturated colours can read as almost black in a genuinely low-light room, losing the very quality they were chosen for. Cool colours of any depth are also difficult, since there is no warm daylight to balance them. And pale cool greys — the most common instinct here — frequently produce exactly the dingy result they were meant to avoid.
Testing advice
Test under the room's actual artificial lighting after dark, not only in daylight, and check the bulbs' colour temperature and colour rendering before deciding — in a room this dependent on artificial light, changing the bulbs can do more for the colour than changing the paint. Test deep and pale options side by side rather than assuming pale is safer.
Lighting simulations and descriptions here are approximations, not exact physical predictions — see methodology.