How STN activity analysis could help to improve DBS stimulation in Parkinson’s disease
- 作者: Sayfulina K.E.1, Gamaleya A.A.2, Sedov A.S.1
-
隶属关系:
- N.N. Semenov Federal Research Center for Chemical Physics, Russian Academy of Sciences
- Burdenko National Scientific and Practical Center for Neurosurgery, Department of Functional Neurosurgery
- 期: 卷 18, 编号 4 (2023)
- 页面: 679-682
- 栏目: Conference proceedings
- ##submission.dateSubmitted##: 15.11.2023
- ##submission.dateAccepted##: 16.11.2023
- ##submission.datePublished##: 15.12.2023
- URL: https://genescells.ru/2313-1829/article/view/623410
- DOI: https://doi.org/10.17816/gc623410
- ID: 623410
如何引用文章
开放存取
##reader.subscriptionAccessGranted##
订阅或者付费存取
详细
Parkinson’s disease (PD) is the second most prevalent neurodegenerative disease globally, associated with the degeneration of dopaminergic neurons in the substantia nigra. The key motor symptoms of PD comprise bradykinesia (voluntary movement slowness and difficulty), rigidity (muscle stiffness), and tremor.
The primary methods to alleviate the extent of indicators in PD are drug therapy using the dopamine precursor levodopa and deep brain stimulation (DBS). The targets for DBS in PD are the nuclei found within the basal ganglia system, specifically the subthalamic nucleus (STN) or the internal segment of the globus pallidus.
There remains an issue of selecting the optimal stimulation program. Identifying the most efficient contact for DBS is a complex and time-consuming process made through clinical observations. One possible way to optimize this process is by analyzing neurophysiological activity of the nucleus and identifying patterns characteristic of the contacts that hold the greatest promise for clinical improvement.
Several studies have investigated the association between LFP and clinical improvements, yet a widely accepted method for choosing stimulation parameters based on STN activity has not been established. The method proposed by Strelow et al. for selecting stimulation contacts based on LFP in the STN is one example [1]. The group’s developed method showed the same efficiency as clinical-driven selection, using only oscillation power in the broad beta range as the activity parameter. Our hypothesis suggests that including more parameters in the analysis could enhance the accuracy of predicting the most effective contacts.
Thus, the study aimed to identify the subthalamic nucleus activity parameters associated with clinical enhancement post-DBS.
The investigation enrolled six patients with PD aged from 44 to 62 years of age (mean 52.8 years, std. 8.2 years) who exhibited akinetic-rigid disease symptoms, namely bradykinesia and rigidity. A total of 12 STN were studied. Implantation of DBS electrodes was carried out in the STN using directional 8-contact electrodes made by St. Jude Medical (USA) with externalization of temporary leads.
LFP recordings were obtained from the implanted electrodes on the first and fifth post-operative days. We examined periods of wakeful rest prior to and following levodopa administration, which we refer to as the OFF-state and ON-state, respectively.
The neurologist assessed clinical symptoms using the UPDRS III scale, evaluating rigidity and bradykinesia on both the left and right sides of the body. The assessment occurred one day prior to surgery and six months after the DBS system was implanted (with consistent stimulation programming during that time). The analysis used the mean improvement in rigidity and bradykinesia. The formula used to calculate the improvement parameter, or the effect of stimulation, is as follows:
(r0–r1)/r0+(g0–g1)/g0)/2.
In this equation, r1 represents the rigidity score after stimulation, r0 represents the preoperative rigidity score, g1 represents the bradykinesia score after stimulation, and g0 represents the preoperative bradykinesia score. All scores were measured during the OFF state.
Recordings were preprocessed using scripts based on the MNE Python software package. We computed 15 bipolar signals using signals from 8 contacts on each electrode. Only those within the stimulation area were used in the analysis.
Spectral analysis was conducted on the bipolar signals, where the power spectral density (PSD) was determined using the Welch method on frequencies ranging from 1–49 Hz with a 1 Hz increment for each bipolar contact. The obtained spectra underwent an aperiodic component subtraction through the fooof method [2], following which the average PSD was obtained for various frequency bands, including theta (4–7 Hz), alpha (8–12 Hz), low (13–19 Hz) and high (20–30 Hz) beta subbands, and the low gamma range (31–49 Hz).
Additionally, bursts within each frequency band were analyzed by extracting periods of increased power spectral density (twice the median value of the recording) from the local field potentials. Various parameters were then calculated for these bursts, including mean burst duration, standard deviation of burst duration, and burst percentage.
Statistical analysis was conducted with R software. The regression analysis aimed to identify parameters linked to clinical improvement. Total improvement in bradykinesia and rigidity served as the dependent variable for the model. Selected LFP parameters for each frequency range and medication state (ON and OFF) were used as predictors in the linear regression model. Factors that did not contribute to model improvement and had high cross-correlations (r >0.8) were excluded to simplify the model.
Analysis uncovered three frequency bands associated with clinical improvement: alpha, low beta, and high beta. A correlation between STN activity parameters and improvement was found in both the ON and OFF states.
The final model (R2=0.55, p <0.001) revealed a noteworthy direct link between clinical improvement and the subsequent factors: decreased beta PSD in the OFF state (p <0.001), and increased alpha PSD in the ON state (p=0.03).
Factors associated with lower clinical improvement included high beta PSD in the ON-state (p=0.018), percentage of high-beta bursts in the ON state (p=0.01), standard deviation of alpha burst duration in the OFF state (p=0.006), and percentage of alpha bursts in the ON state (p <0.001).
Low beta-band power’s association with clinical improvement is consistent with prior literature [3], and beta burst activity parameters have previously been linked to improvement [4]. However, we discovered differential associations with improvement for both low and high beta oscillations, and an additional association with improvement for alpha activity. Based on our findings, we suggest that several STN LFP parameters beyond broad beta power, such as bursts and PSD within alpha, low-beta, and high-beta bands, are associated with clinical improvement. These parameters could be used to determine the most effective stimulation contacts. The evaluation of these parameters requires further investigation.
全文:
Parkinson’s disease (PD) is the second most prevalent neurodegenerative disease globally, associated with the degeneration of dopaminergic neurons in the substantia nigra. The key motor symptoms of PD comprise bradykinesia (voluntary movement slowness and difficulty), rigidity (muscle stiffness), and tremor.
The primary methods to alleviate the extent of indicators in PD are drug therapy using the dopamine precursor levodopa and deep brain stimulation (DBS). The targets for DBS in PD are the nuclei found within the basal ganglia system, specifically the subthalamic nucleus (STN) or the internal segment of the globus pallidus.
There remains an issue of selecting the optimal stimulation program. Identifying the most efficient contact for DBS is a complex and time-consuming process made through clinical observations. One possible way to optimize this process is by analyzing neurophysiological activity of the nucleus and identifying patterns characteristic of the contacts that hold the greatest promise for clinical improvement.
Several studies have investigated the association between LFP and clinical improvements, yet a widely accepted method for choosing stimulation parameters based on STN activity has not been established. The method proposed by Strelow et al. for selecting stimulation contacts based on LFP in the STN is one example [1]. The group’s developed method showed the same efficiency as clinical-driven selection, using only oscillation power in the broad beta range as the activity parameter. Our hypothesis suggests that including more parameters in the analysis could enhance the accuracy of predicting the most effective contacts.
Thus, the study aimed to identify the subthalamic nucleus activity parameters associated with clinical enhancement post-DBS.
The investigation enrolled six patients with PD aged from 44 to 62 years of age (mean 52.8 years, std. 8.2 years) who exhibited akinetic-rigid disease symptoms, namely bradykinesia and rigidity. A total of 12 STN were studied. Implantation of DBS electrodes was carried out in the STN using directional 8-contact electrodes made by St. Jude Medical (USA) with externalization of temporary leads.
LFP recordings were obtained from the implanted electrodes on the first and fifth post-operative days. We examined periods of wakeful rest prior to and following levodopa administration, which we refer to as the OFF-state and ON-state, respectively.
The neurologist assessed clinical symptoms using the UPDRS III scale, evaluating rigidity and bradykinesia on both the left and right sides of the body. The assessment occurred one day prior to surgery and six months after the DBS system was implanted (with consistent stimulation programming during that time). The analysis used the mean improvement in rigidity and bradykinesia. The formula used to calculate the improvement parameter, or the effect of stimulation, is as follows:
(r0–r1)/r0+(g0–g1)/g0)/2.
In this equation, r1 represents the rigidity score after stimulation, r0 represents the preoperative rigidity score, g1 represents the bradykinesia score after stimulation, and g0 represents the preoperative bradykinesia score. All scores were measured during the OFF state.
Recordings were preprocessed using scripts based on the MNE Python software package. We computed 15 bipolar signals using signals from 8 contacts on each electrode. Only those within the stimulation area were used in the analysis.
Spectral analysis was conducted on the bipolar signals, where the power spectral density (PSD) was determined using the Welch method on frequencies ranging from 1–49 Hz with a 1 Hz increment for each bipolar contact. The obtained spectra underwent an aperiodic component subtraction through the fooof method [2], following which the average PSD was obtained for various frequency bands, including theta (4–7 Hz), alpha (8–12 Hz), low (13–19 Hz) and high (20–30 Hz) beta subbands, and the low gamma range (31–49 Hz).
Additionally, bursts within each frequency band were analyzed by extracting periods of increased power spectral density (twice the median value of the recording) from the local field potentials. Various parameters were then calculated for these bursts, including mean burst duration, standard deviation of burst duration, and burst percentage.
Statistical analysis was conducted with R software. The regression analysis aimed to identify parameters linked to clinical improvement. Total improvement in bradykinesia and rigidity served as the dependent variable for the model. Selected LFP parameters for each frequency range and medication state (ON and OFF) were used as predictors in the linear regression model. Factors that did not contribute to model improvement and had high cross-correlations (r >0.8) were excluded to simplify the model.
Analysis uncovered three frequency bands associated with clinical improvement: alpha, low beta, and high beta. A correlation between STN activity parameters and improvement was found in both the ON and OFF states.
The final model (R2=0.55, p <0.001) revealed a noteworthy direct link between clinical improvement and the subsequent factors: decreased beta PSD in the OFF state (p <0.001), and increased alpha PSD in the ON state (p=0.03).
Factors associated with lower clinical improvement included high beta PSD in the ON-state (p=0.018), percentage of high-beta bursts in the ON state (p=0.01), standard deviation of alpha burst duration in the OFF state (p=0.006), and percentage of alpha bursts in the ON state (p <0.001).
Low beta-band power’s association with clinical improvement is consistent with prior literature [3], and beta burst activity parameters have previously been linked to improvement [4]. However, we discovered differential associations with improvement for both low and high beta oscillations, and an additional association with improvement for alpha activity. Based on our findings, we suggest that several STN LFP parameters beyond broad beta power, such as bursts and PSD within alpha, low-beta, and high-beta bands, are associated with clinical improvement. These parameters could be used to determine the most effective stimulation contacts. The evaluation of these parameters requires further investigation.
ADDITIONAL INFORMATION
Funding sources. This work was supported by grant No. 22-15-00344 from the Russian Science Foundation.
作者简介
K. Sayfulina
N.N. Semenov Federal Research Center for Chemical Physics, Russian Academy of Sciences
编辑信件的主要联系方式.
Email: kseniasayfulina@gmail.com
俄罗斯联邦, Moscow
A. Gamaleya
Burdenko National Scientific and Practical Center for Neurosurgery, Department of Functional Neurosurgery
Email: kseniasayfulina@gmail.com
俄罗斯联邦, Moscow
A. Sedov
N.N. Semenov Federal Research Center for Chemical Physics, Russian Academy of Sciences
Email: kseniasayfulina@gmail.com
俄罗斯联邦, Moscow
参考
- Strelow JN, Dembek TA, Baldermann JC, et al. Local Field Potential-Guided Contact Selection Using Chronically Implanted Sensing Devices for Deep Brain Stimulation in Parkinson’s Disease. Brain Sciences. 2022;12(12):1726. doi: 10.3390/brainsci12121726
- Donoghue T, Haller M, Peterson EJ, et al. Parameterizing neural power spectra into periodic and aperiodic components. Nature Neuroscience. 2020;23(12):1655–1665. doi: 10.1038/s41593-020-00744-x
- Feldmann LK, Lofredi R, Neumann WJ, et al. Toward therapeutic electrophysiology: beta-band suppression as a biomarker in chronic local field potential recordings. NPJ Parkinson’s Disease. 2022;8(1):44. doi: 10.1038/s41531-022-00301-2
- Tinkhauser G, Pogosyan A, Little S, et al. The modulatory effect of adaptive deep brain stimulation on beta bursts in Parkinson’s disease. Brain. 2017;140(4):1053–1067. doi: 10.1093/brain/awx010
补充文件
