Synchronized Intracranial Electrical Activity and Gait Recording in Parkinson’s Disease Patients With Freezing of Gait

Abstract

BackgroundThis study aimed to describe a synchronized intracranial electroencephalogram (EEG) recording and motion capture system, which was designed to explore the neural dynamics during walking of Parkinson’s disease (PD) patients with freezing of gait (FOG). Preliminary analysis was performed to test the reliability of this system.MethodsA total of 8 patients were enrolled in the study. All patients underwent bilateral STN-DBS surgery and were implanted with a right subdural electrode covering premotor and motor area. Synchronized electrophysiological and gait data were collected using the Nihon Kohden EEG amplifier and Codamotion system when subjects performed the Timed Up and Go (TUG) test. To verify the reliability of the acquisition system and data quality, we calculated and compared the FOG index between freezing and non-freezing periods during walking. For electrophysiological data, we first manually reviewed the scaled (five levels) quality during waking. Spectra comprising broadband electrocorticography (ECoG) and local field potential (LFP) were also compared between the FOG and non-FOG states. Lastly, connectivity analysis using coherence between cortical and STN electrodes were conducted. In addition, we also use machine learning approaches to classified FOG and non-FOG.ResultsA total of 8 patients completed 41 walking tests, 30 of which had frozen episodes, and 21 of the 30 raw data were level 1 or 2 in quality (70%). The mean ± SD walking time for the TUG test was 85.94 ± 47.68 s (range: 38 to 190.14 s); the mean ± SD freezing duration was 12.25 ± 7.35 s (range: 1.71 to 27.50 s). The FOG index significantly increased during the manually labeled FOG period (P < 0.05). The beta power of STN LFP in the FOG period was significantly higher than that in the non-FOG period (P < 0.05), while the band power of ECoG did not exhibit a significant difference between walking states. The coherence between the ECoG and STN LFP was significantly greater in high beta and gamma bands during the FOG period compared with the shuffled surrogates (P < 0.05). Lastly, STN-LFP band power features showed above-chance performance (p < 0.01, permutation test) in identifying FOG epochs.ConclusionIn this study, we established and verified the synchronized ECoG/LFP and gait recording system in PD patients with FOG. Further neural substrates underlying FOG could be explored using the current system.

Machine learning for adaptive deep brain stimulation in Parkinson’s disease: closing the loop

“…These studies used, respectively, K-nearest neighbors and support vector machines methods to predict motor symptom severity and Freezing of Gait (FoG) events. Liu et al took advantage of yet another RF model to classify FoG states based on the Codamotion 3-D movement capture system and synchronized intracranial EEG data, further broadening the understanding of movement dynamics and contributing to bridging the gap between peripheral and intracranial signals [ 48 ].…”

Fronto‐parieto‐subthalamic activity decodes motor status in Parkinson's disease

“…All participants provided written informed consent to participate, which was approved by the Institutional Review Board of Beijing Tiantan Hospital. The subjects' preoperative assessments can be found in our previous reports 13–15 . The participants' demographic and clinical information is summarized in Table 1.…”

“…The subjects' preoperative assessments can be found in our previous reports. 13 , 14 , 15 The participants' demographic and clinical information is summarized in Table 1 .…”

Artificial intelligence in medicine

“…The recent advance in deep brain stimulation allows the delivery of a specific electric energy program in a very restricted anatomical area. In addition, this new system can recognize pathologic patterns of local field potentials and respond with a specific schedule of electric pulse trains 7 .…”