The concept of adaptive DBS (aDBS) in Parkinson’s disease (PD) is based on the use of local field potentials (LFPs) as control variable to update DBS parameters. Despite some controversial results on the specific clinical measures correlating with basal ganglia LFP rhythms, present evidence supports the hypothesis that LFP activity changes in response to the patient’s clinical state, movement execution, and also cognitive and behavioural stimuli. The most studied and debated LFP oscillations are beta frequencies (13–35 Hz). Present results show that beta frequencies reflect the patient’s motor state: beta-band LFP oscillations are disrupted by dopaminergic medication, correlate with movement preparation and execution, and with motor representation, including action organization and context recognition. Non-linear analysis evidenced also that the lower portion of the beta band, without dopaminergic medication, generates harmonic components and non-linear synchronizations with other LFP rhythms that are thought to be partly responsible of the patient’s motor symptoms. Despite not present in all the patients (only the 70% of PD patients show a clear beta peak), the strong correlation between the beta rhythm and patient’s motor state made it a good candidate for the first aDBS experiments, in which DBS intensity is adapted according to beta band changes.