Have a look on YouTube for some of the videos about the design & development of the class 91. The traction motors on them are pretty much in the last place you’d expect, mainly due to a desire not to repeat the mistakes of the 86s and the way they knocked lumps out of the track.
I was under the impression that the main design challenge for the Class 91 was that the motors are so bloomin' large!
To simplify some very long stories:
Class 81-85 had various forms of flexible drive so the motor was sitting on the sprung part of the bogie but could still mesh with the gears on the axle despite the relative movement between the two. This was eliminated on the class 86, no doubt to save money. They had nose-suspended traction motors where one end of the motor effectively rested on the axle so there was no relative movement to be taken up. That meant about half the motor mass was unsprung, imposing high transient loads on the track which led to problems such as clay pumping through the ballast. For that reason the 87s had a more flexible drive arrangement - I think a hollow motor armature with another shaft through the middle linking to the gears, with the two being connected by a bush at the non-gear end so the motor could move around on the springs and the gears would still mesh. The 86s were either rebuilt to reduce the problem or restricted to lower speeds. In all these the motor was positioned crossways so the armature was parallel to the axle.
Something similar was done for the HST but with the 91 being specified for 140mph and presumably needing bigger motors for more power a different approach was needed. I believe the APT had put the motors lengthways within the body with cardan shafts down to right-angle gears on each axle. The 91 kept the same basic arrangement but effectively hung the motors below the body so they were within the bogie frame but not attached to it. Pendolinos and Voyagers have (I think) a motor lengthways near each end of the underfloor housing, with a cardan shaft to the adjacent axle.