Daily activities like preparing a meal rely on the ability to arrange thoughts and actions in the right order. Patients with Parkinson’s disease (PD) have difficulties in sequencing tasks. Their deficits in sequential working memory have been associated with basal ganglia dysfunction. Here we demonstrate that altered parietal alpha-theta oscillations correlate with sequential working memory in PD. We included 15 PD patients (6 women, mean age 66.0 years), 24 healthy young (HY, 14 women, mean age 24.1 years), and 16 older participants (HO, 7 women, mean age 68.6 years). Participants completed a picture ordering task with scalp EEG recording, where they arranged five pictures in a specific order and memorized them over a delay. During encoding and the delay, the baseline alpha peak frequency for ordered trials was 10.0 Hz in HY, 10.1 Hz in HO, and 8.9 Hz in PD. PD showed a lower baseline alpha peak frequency with higher alpha power than HY and HO. PD with a higher baseline alpha power tended to respond more slowly. In response to sequence manipulation, the frequency of maximal power change between random and ordered trials (F_max) was 10.0 Hz in HY, 11.3 Hz in HO, and 7.7 Hz in PD. PD showed a lower F_max than HY and HO, whereas HO showed a higher F_max than HY. Compared to PD with F_max in the alpha band (8-15 Hz, n=11), PD with F_max in the theta band (4-7 Hz, n=4) tended to show a higher ordering-related accuracy cost in the picture ordering task and lower accuracy and slower responses in an independent working memory task. In conclusion, altered baseline alpha oscillations and task-dependent modulation of alpha-theta oscillations may be a neural marker of poor sequential working memory in PD. 

Sample: 24 healthy young participants, 16 healthy older participants, 15 patients suffering from idiopathic Parkinson's Disease.

Recording: EEG data were recorded from 29 tin electrodes mounted on an elastic cap using an BrainAmp amplifier (Brain Products GmbH, Gilching, Germany). The electrodes were placed according to the international 10-20 system. Two additional electrodes were placed on the bilateral mastoids. The vertical electrooculogram (EOG) was recorded from electrodes above and below the left eye. The horizontal EOG was recorded from electrodes on the outer canthi of each eye. The data were sampled at 250 Hz, referenced online against the right mastoid, and filtered with a bandpass filter of 0.016 Hz to 1kHz. Electrode impedances were kept below 5 kΩ. 

Preprocessing: EEG data were preprocessed with the EEGLAB toolbox. The data were filtered with a low-pass filter of 48 Hz, re-referenced to the mean signal of the bilateral mastoids, and segmented into EEG epochs encompassing the entire trial ([-2.5 20] s around the trial onset). The epoched data were subjected to an independent component analysis algorithm to remove eye movement and other artifacts. The ICA corrected epochs were then visually inspected, and artifact afflicted epochs excluded from further analysis. EEG epochs were baseline-corrected by subtracting the mean voltage before the trial onset. Approximate 24% of ordered trials and 31% of random trials were rejected as artifacts. There was no group difference in the number of artifact-free trials. For the subsequent time frequency analyses data were exported to the Fieldtrip format. 

For group assignment see file subject_group.txt

For task description and meaning of trigger codes see file subject_group.txt