Concur with Justin, but to expand on info a bit:
Going up in voltage is OK, but 50 to 400? It's a bit much. 22u 50V is a fairly small device, whereas 22u 400V is huge in comparison (doing a real comparison shows that the 400V cap is twice the diameter, twice the height and twice the pin pitch). A more sensible step-up would be to 22u 63V, where the can diameter would be a bit bigger (by a mm) and the height greater by a couple of mm. Pin pitch may well be the same or just slightly bigger by 0.5mm to 1mm..... the 22u 63V part will likely be a neat replacement for the 22u 50V part and not look like a shoved in amateur bodge.
Understanding the circuit is a big benefit when altering cap values.
Example #1, power rail decoupling for op-amps:
16V caps may already be fitted - this is a bit close to the limits as the power rails on an op-amp are typically +/-15V, but if higher (to give greater audio headroom) at say, +/-16V to +/-18V, the 16V caps will be at their limit. In this situation it is well worthwhile replacing the 16V caps for 25V caps. There is nothing to be gained by putting in caps of a higher voltage.
Regarding the uF value, again, a good understanding of the circuit is an advantage when considering increasing the cap's value.
Example #2, audio coupling capacitors:
Staying with op-amps and +/-17V power rails, fitting caps of greater than 25V for audio coupling is a waste of time...... In many cases you can fit caps of only 4V - remember that the voltage rating of a cap is the DC voltage across the terminals, and an op-amp circuit with more than 4V of DC offset is probably very unwell. In the real world, DC offsets in the mV range are typical. One approach is to use low-voltage caps (say, 10V) for inter-stage coupling within a piece of equipment, and then to use higher voltage caps (say, 25V) for coupling to external equipment. This approach is based on the idea that you can never be sure that there will not be a large DC voltage on the end of an audio cable. The classic example of high levels of DC across an audio coupling cap is a microphone input with 48V phantom power, and in these circuits you WILL see caps rated at 50V or 63V. (Personal experience has shown that some surface-mount 50V caps are too close to the limits and fail).
Inter-stage coupling capacitors can have their values tweaked, but again I must stress that understanding the circuit is important.
Changing from "standard" electrolytics to low ESR is also possible, but understanding the circuit is important so as to know where a low ESR cap will yield its best performance, or indeed improve the performance of the circuit.
