When using the polythrone is formed a lot of foam as the speed of tissue destruction in it is too high. Usual homogenizing at normal cooling ("Braun") (it is better to do everything in the cold room) allows to receive both good and stable brain synaptosomes and fine reproduced activities of enzymes.
If you're using an automated homogenizer (rotor-stator, ultrasonic, bead mill, etc.) you're going to get denaturation. It's practically unavoidable. If your analysis depends on retaining native state proteins, you'll probably want to stick with a manual dounce homogenizer. This will minimize foaming, cavitation, vortexing, and creation of liquid-air interfaces more generally.
If you're not worried about the proteins being in the native state, any type of homogenization can be used. Brain tissue homogenizes very easily. The consideration should then be based more on sample size, throughput, and projected future needs.
If you would like, you can learn more about different homogenization technologies here: http://homogenizers.net/pages/ac-application-center