Fig. 3: Case study of out-of-clinic data.

a Representation of Timeline and Events of the selected hemisphere. The recordings cover approximately one month in a conventional DBS protocol (continuous fixed stimulation amplitude). b Estimated circadian profile of the Timeline recording in (a). The LFP power is higher during waking hours. c Fast Fourier Transform (FFT) profiles for each of the recorded Events. The observed median profiles were similar, with the highest variation around the 15 Hz peak. d Daily distribution of ‘Medication’ and ‘I feel stuck’ episodes. Notice that distributions are similar, with three periods of higher incidence (6 h, 11-12 h, 20-21 h). e Distribution of ‘I feel stuck’ events relative to ‘Medication’ events (triggered at zero). Interestingly, ‘I feel stuck’ episodes are mostly reported after ‘Medication’. f Event-triggered average plots showing the power of the selected band in the proximity of ‘Medication’ and ‘I feel stuck’ episodes. The mean power decreases after the two events.