Hi Could you please elaborate your question? Are you doing experiments with rodents? If for rodents, I guess we inject to the third ventricle because it is easy to access, it is located relatively close to skull surface, and has a large volume. We can easily use bregma and lamda as landmarks to target the specific injection location. 

I lack direct experience with 6OHDA, but in general the choice of lateral, ventricle, third ventricle, intracisternal or intraparenchymal adminsitration depends firstly on the proximity/accessibility to the target system you want to affect, and secondly, on the volume you need to administer. Please check these matters with authors experienced in studies similar to yours.