Galactic cosmic rays (GCR) consist of high-energy particles that originate far outside solar system and reach the Earth. Studies show that for every decrease in GCR intensity there is an increase in interplanetary magnetic field (IMF). Interplanetary magnetic field can be increased due to different phenomena in solar atmosphere, such as coronal mass ejections (CMEs) – phenomena driven by the energy released from magnetic field. Furthermore, for every increase in IMF there is a decrease in GCR intensity. Such a short-term depression in the GCR count is called Forbush decrease (FDs) and it is named after the American physicist Scott E. Forbush who discovered it in 1937. Forbush decreases can be observed by particle detectors several days after the CME. Forbush decreases caused by fast coronal mass ejections (CMEs) often show a two-step decrease where the first step is attributed to the shock/sheath region, while the second step is attributed to the closed magnetic structure. Since the difference in size of shock and sheath region is significant, and since there are observed effects that can be related to shocks and not necessarily to the sheath region we expect that the physical mechanisms governing the interaction with GCRs in these two regions are different. We therefore aim to analyse interaction of GCRs with heliospheric shocks only. First we study the role of two physical mechanisms that could directly affect the galactic cosmic ray (GCR) count rate in the shock-sheath region ahead of the CME. These are the magnetic mirror effect at the shock front and the magnetic field compression behind the shock. The effects of the mentioned two mechanisms for the shock-related phase of the FD are quantified by employing a simplified magnetic configuration that includes an oblique magnetohydrodynamical (MHD) fast-mode shock. Using the jump relations at the shock we determine the fraction of GCR particles that are mirrored at the shock, as well as the increase of the unit-area flux of the transmitted particles due to the compression of the magnetic field in the downstream region. A combination of these two effects results in a decrease of the GCR count rate. The calculated FD amplitudes attain values up to 45 %, which is much larger than are the observed amplitudes. This confirms that the magnetic mirror approximation is not a valid one which is already indicated by the fact that the shock thickness is much smaller than the gyro-radius of the GCR particle (proton). In the next step we approximate the shock by a structure where the magnetic field linearly changes with position within this structure. We assume protons of different energy, different pitch angle and different incoming direction and also vary the shock parameters such as the magnetic field strength and orientation, as well as the shock thickness. We solve the set of three differential equations to obtain proton trajectories for different initial speed components, i.e. different initial conditions. The results are in an agreement with observations and demonstrate that protons with higher energies are less likely to be reflected. Also, thicker shocks and shocks iv with stronger field reflect protons more efficiently. In the second part of this research we focus on the magnetic structure of the CME. We use the observationally obtained FD amplitude to determine the flux rope (FR) expansion parameters. We use the diffusion-expansion model, ForbMod (Dumbović et al., 2018). This model is based on the assumption that the FR is a long expanding cylinder and GCR particles enter the FR via perpendicular diffusion. In the original model it is assumed that all particles have the same energy, i.e. rigidity of 1 GV which corresponds to the most common energy in energy spectrum. Parameters of expansion can be analytically expressed according to Eq. (1) in Rodari et al. (2018). We compare the values obtained by this simple model with the values obtained numerically from the same model adjusted to include particles with energies ranging from 50 MeV to 100 GeV. In that case expansion parameters can not be expressed analytically and have to be calculated numerically. The values obtained both ways are then used to determine axial magnetic flux near the Sun. These values of magnetic flux are compared to the values obtained using Eq. (6) in Scolini et al., (2020). They use an independent method based on the connection between CME kinematics and magnetic reconnection during and immediately after the eruption. Results indicate that ForbMod model is not applicable for the events where the expansion speed is small or even negative. For these events, magnetic fluxes were several orders of magnitude larger than expected. The smallest discrepancies in results are found in the events where the power-law index of magnetic flux expansion are large.