Context : Recent spectroscopic and photometric surveys are providing a comprehensive view of the Milky Way bulge stellar population properties with unprecedented accuracy .
This in turn allows us to explore the correlation between kinematics and stellar density distribution , crucial to constraint the models of Galactic bulge formation .

Aims : The Giraffe Inner Bulge Survey ( GIBS ) revealed the presence of a velocity dispersion peak in the central few degrees of the Galaxy by consistently measuring high velocity dispersion in three central most fields .
Due to suboptimal distribution of these fields , all being at negative latitudes and close to each other , the shape and extension of the sigma peak is poorly constrained .
In this study we address this by adding new observations distributed more uniformly and in particular including fields at positive latitudes that were missing in GIBS .

Methods : MUSE observations were collected in four fields at ( l,b ) = ( 0 ^ { \circ } , +2 ^ { \circ } ) , ( 0 ^ { \circ } , -2 ^ { \circ } ) , ( +1 ^ { \circ } , -1 ^ { \circ } ) , and ( -1 ^ { \circ } , +2 ^ { \circ } ) .
Individual stellar spectra were extracted for a number of stars comprised between \sim 500 and \sim 1200 , depending on the seeing and the exposure time .
Velocity measurements are done by cross-correlating observed stellar spectra in the CaT region with a synthetic template , and velocity errors obtained through Monte Carlo simulations , cross-correlating synthetic spectra with a range of different metallicities and different noise characteristics .

Results : We measure the central velocity dispersion peak within a projected distance from the Galactic center of \sim 280 pc , reaching \sigma V _ { GC } \sim 140 km/s at b=-1 ^ { \circ } .
This is in agreement with the results obtained previously by GIBS at negative longitude .
The central sigma peak is symmetric with respect to the Galactic plane , with a longitude extension at least as narrow as predicted by GIBS .

As a result of the Monte Carlo simulations we present analytical equations for the radial velocity measurement error as a function of metallicity and signal-to-noise ratio for giant and dwarf stars .

Conclusions :