I would not be surprised if comfort and performance peak at different temperatures. A non competitive cyclist might be inclined to ride harder when comfortable, but that's not the same as saying performance is maximized when comfortable. I tend to believe peak performance occurs at a lower temperature than peak comfort for most of us. I think the energy cost of heat rejection and the negative impact of to much heat are both high, and occur below most peoples' comfort level.
The article mentioned clothing changes, which to me is an uncontrolled variable in the experiment. I don't think their data is valid.