We study a two-component Fermi gas that is so strongly polarized that it remains normal fluid at zero temperature. We calculate the occupation numbers within the particle-particle random-phase approximation, which is similar to the Nozières-Schmitt-Rink approach. We show that the Luttinger theorem is fulfilled in this approach. We also study the change of the chemical potentials which allows us to extract, in the limit of extreme polarization, the polaron energy. [3]. In this theory, only equal densities of the two species forming the pairs (which we will denote by spin indices σ =↑, ↓) were considered. Nowadays, the crossover can be realized in experiments with ultracold trapped atoms whose scattering length a can be tuned with the help of a Feshbach resonance [4]. In these experiments, it is also possible to study systems with different densities of the two species, n ↑ = n ↓ [5]. In this way, one tries to discover, e.g., phases with exotic pairing, like the Fulde-Ferrel-Larkin-Ovchinnikov (FFLO) phase [6,7]. Also the extremely polarized case, where only a single particle of spin ↓ is put into a Fermi sea of spin ↑, bears interesting physics: when passing though the Feshbach resonance from the attractive (a < 0) to the repulsive (a > 0) side, one expects that the ground state transforms from a Fermi sea plus a fermionic quasiparticle, the so-called polaron, into a Fermi sea plus a bosonic molecule [8]. To our knowledge, there is so far no unique many-body theory that can describe the imbalanced Fermi gas in the cross-over and that reproduces in the limit of extreme polarization the polaronic and the molecular ground state, depending on the value of the scattering length a.As mentioned before, the original NSR theory was formulated in order to describe the BEC-BCS crossover in a two-component (σ =↑, ↓) Fermi gas with equal populations. Within this approach, the critical temperature T c as a function of the chemical potential µ is obtained from the Thouless criterion, i.e., it is the temperature where the in-medium T matrix develops a pole,where ω and k are, respectively, the total energy (measured from 2µ) and momentum of the pair, and Γ is ob- tained by summing ladder diagrams, see Fig. 1. As a function of µ, the critical temperature obtained in this way is exactly the same as within BCS theory. The difference between BCS and NSR comes from the inclusion of pair correlations into the relationship n(µ) between the number density and the chemical potential. This is done by including diagrams of the type shown in Fig. 2(a) into the thermodynamic potential Ω(µ, T ) and then computing the density from n = −∂Ω/∂µ. This is equivalent to calculating the density from [9]where ω n = nπT is a fermionic (n odd) Matsubara frequency and G is the single-particle (s.p.) Green's function with at most one self-energy insertion,