The importance of the middle atmosphere for the weather and climate on Earth is increasingly realized. Variability and feedback processes in the middle atmosphere need to be better understood and form the subject of this thesis. Initially, the focus has been on the variability of the summer polar mesopause, which is the coldest place in the Earth's system. The variability of this region is driven by a variety of atmospheric processes, such as atmospheric waves and the solar cycle and is even coupled to the atmosphere on other side of the globe through interhemispheric coupling. The low temperatures in the summer polar mesopause allow for thin ice clouds to form: noctilucent clouds (NLCs). It is investigated how well the Canadian Middle Atmosphere Model (CMAM30), in which the NLCs are represented in terms of a simple model, can be used to study zonal mean NLC variability.  Comparing to satellite data, it is shown that the basic NLC characteristics, such as seasonal onsets and development, interannual variability and interhemispheric differences, are well captured by the model. The role of the winter residual circulation in shaping the conditions of the summer polar mesopause is also investigated, using the Whole Atmosphere Community Climate Model (WACCM). It is found that without the gravity waves in winter, the summer mesopause region would be significantly warmer. This means that the interhemispheric coupling mechanism has a net cooling effect on the summer mesopause regions. In addition, the effect of the solar cycle on the summer polar mesopause is studied. In CMAM30, there is no substantial temperature change due to the solar cycle. It is shown that there is an enhanced circulation in this region during solar maximum as compared to solar minimum, which causes adiabatic cooling counteracting the direct effect of the solar cycle. Finally, feedbacks in the middle atmosphere are studied using WACCM. The Climate Feedback Response Analysis Method (CFRAM) is used to examine the middle atmosphere response to a doubling of the CO₂-concentration with respect the pre-industrial state. It was found that the temperature response to direct CO₂ forcing would be approximately -9 K in the middle atmosphere. This cooling is being mitigated by the combined effect of the different feedbacks processes, the strongest of which being the ozone feedback. The dynamical feedback has large effects on the temperatures locally, while the role of the cloud, albedo and water vapor feedback are small in the middle atmosphere.