The outcome of the cycloaddition between activated ketenes and alpha,beta-unsaturated imines has been investigated both experimentally and theoretically. Our results indicate that activated monosubstituted ketenes yield exclusively [2 + 2] cycloadducts. Disubstituted activated ketenes yield [2 + 2] and/or [4 + 2] cycloadducts. In one case, an unexpected piperidin-2-one has been obtained, although its relative abundance with respect to the corresponding [2 + 2] or [4 + 2] cycloadducts can be minimized by the proper choice of experimental conditions. The ability of different ab initio and semiempirical methods to account for these results has been tested. The best agreement between theory and experiment is achieved at the MP2/6-31G level of theory, with solvent effects taken into account. The semiempirical hamiltonian AM1, at the RHF level, tends to overestimate the stability of the transition structures leading to six-membered cycloadducts, whereas 3 x 3CI-HE/AM1 and CASSCF(2,2)/6-31G methods tend to overestimate the stability and the biradical character of the transition structures leading to [2 + 2] cycloadducts.