Considerable efforts are currently being devoted to enhance the speed, spatial resolution and the size of the 3D sample volumes in which calcium imaging methods can capture neuronal network activity in different model systems. In the mammalian brain, tissue scattering severely limits the use of parallel acquisition techniques such as wide-field imaging and, as a consequence, methods based on two-photon point-scanning (2PM) have become the method of choice. However, 2PM faces severe restrictions due to technical limitations such as scan speed, laser power, and those related to the fluorescent probes, calling for conceptually new approaches to enhance the performance of two-photon calcium imaging schemes. Here we provide a detailed quantitative evaluation and comparison of different excitation/detection modalities from the perspective of detecting neuronal activity that are based on different point-spread functions (PSF), laser repetition rates and sampling strategies. We demonstrate the conditions for which imaging speed and signal-to-noise ratio are optimized for a given average power. Our results are based on numerical simulations which are informed by experimentally measured parameters and show that volumetric field of view and acquisition speed can be considerably improved compared to traditional 2PM schemes by a holistic optimization approach.