This is probably more obvious to most people than it was to me. My judgment is distorted by a lifetime of urban bicycling.
It was very difficult to teach #1 and #2 to ride a bicycle (it would be easier today - we know more). Almost as hard as teaching them to swim. They did well in the end though. #1 competed in high school mountain biking and I think he is a relatively safe urban cyclist. His impulsivity and rigidity are balanced by native caution and seemingly strong visual processing.
#2 has a substantially higher IQ than #1, but he’s a weaker bicyclist. We did a trial bike ride to school today; he did well with guidance but he was exhausted. I think the relatively simple ride was cognitively draining. #2 is closer to the classic Asperger’s pattern — persistent attention to the external world is very difficult. He may never be able to bicycle commute safely, though he does well mountain biking (and inline skating - remarkable balance there).
In retrospect I’m not sure urban bicycle commuting is cognitively less demanding than urban driving. There’s more time to plan actions, but there’s a lot more judgment involved. By comparison car driving is more rule-bound.
Special needs mountain/gravel biking, or bicycling on separated trails, works for #2. Bicycling on city streets - not so much.