I believe it does not matter whether you are brakeing or acceleration. I used to think it did, but recently I had some things explained to me and did some thinking and drawing out of my own.
Ok, I'll admit now, I am not good at relaying these explainations without being able to actively draw it all out for you as I speak it.
But here goes,
The reason both accelerating and decelerating behave in the same manner will be explained
Consider a front wheel as it sits in a stable state. The weight of the car is balanced by the upward force of the suspension. If we set no toe at that point both wheels face directly forward. If the car accelerates hard the tyre moves down (hence the car body moves upwards). If toe were to stay the same the steering arms would need to self-lengthen, but obviously they don't as they are solid. So if we say the distance from steering rack to the ball joint (where the arms/rods that protrude from the rack it self with the rubber boot on them can move) is "r". This is the radius of an arc in which the balljoint will travel since "r" is constant. In order to keep "r" constant the wheel must turn in order to move up and down. Whether up and down the same effect is achieved, the rear of the wheel (since the rack is rear mounted) will turn towards the car.
Now consider decelerating - in which the tyres moves upwards. If we decelerate as 'hard' as we accelerated earlier the same toe out should occur.
In the pic (which is dodgy, sorry, I only got paint!) the blue point, is the point in which the steering rod/arm can rotate around. This is where the rubber boot on the steering rack exists. The green bar is the steering rod/arm/whatever you call it, which joins the steering arm at the balljoint which is black. The ball joint, towards the wheel is solid.
Anyway it should be seen that the same effect occurs either direction of wheel travel - up or down.
The red line is the arc in which the ball joint will travel. Also, the travel of the wheel is along a designated path, up and down. If the wheel can only travel directly up and down, and the steering rod travels in an arc due to having a fixed point at the steering rack end, it should be seen that something must give. This something that 'gives' is the toe angle of the wheel. The rear of the swivel hub moving inwards as the wheel goes up (the further up the more angle) and as the wheel goes down (as above).
So if you intend to have you wheel at any extreme often, either up or down, toe in to the required amount is what you would want to ensure the straightest possible alignment of the front wheels under acceleration.
Ok, I am not good at explaining things, but I had a go! Yay.
BTW, I've noticed people too often just think of FWD cars being opposite to RWD cars, people need to think of a Mini, which happens to be FWD, study the Mini's suspension geometry as a seperate entity using knowledge that is applicable to all types of cars. - Take the question on it's own merits! It also helps to draw out some points and some lines, then think of various circumstances in which the object you are considering behaves.
Cheers, Tim.
_________________
1964 Morris 850, 1330 Supercharged - 81.8hp atws.
1975 Leyland Mini S 1100S powered - Nice and reliable.
1977 Leyland Mini LS - Project LS-T
