Recently, we demonstrated the prevalent role of cerebellar networks in the
acquisition of the procedural components of spatial information by testing
hemicerebellectomized (HCbed) rats in a classical spatial task, the Morris
water maze (MWM). As procedures used in the water maze are a mixture of dif
ferent components (that is, general procedures, exploration procedures, dir
ect reaching procedures), for optimally solving a spatial task all procedur
al components must be opportunely managed. Thus, severely impaired procedur
al learning of cerebellar origin can be better comprehended by fractionatin
g the procedural facets. To this aim, a two-step water-maze paradigm was em
ployed. Normal rats were first trained to search for a hidden platform move
d to a different position in each trial, utilizing a water maze setting in
which visual cues were abolished by heavy black curtains surrounding the ta
nk. In this paradigm, normal animals solved the task by using general and e
xploration procedures, but they could not use direct reaching skills. A sub
group of these pretrained animals was then HC-bed and, after recovery from
cerebellar lesion, was tested in a water maze with normal environmental cue
s available, a paradigm in which normal animals develop abilities for reach
ing the target with very direct trajectories. Pretrained HCbed animals, how
ever did not display the typical spatial deficits of naive HCbed rats, pers
isted in exhibiting the scanning strategy learned during pretraining, and n
ever displayed direct reaching skills. In conclusion, cerebellar networks a
ppear to be involved in the acquisition of all procedural facets necessary
for shifting behavior within the maze until direct reaching of the platform
. The lack of flexibility in changing exploration strategies displayed by p
retrained HCbed rats is interpreted by taking into account the well-known c
erebellar frontal interplay sculpting a specific cerebellar role in the acq
uisition of spatial procedural steps.