Deep brain stimulation (DBS) is increasingly used in pediatric dystonia, facilitated by published consensus programming guidelines, 1 yet limited data exists to guide programming or presurgical decision‐making. DBS can result in remarkable improvements for some pediatric dystonia patients while others experience little benefit. 2 Several factors may contribute to this variability, including the spectrum of genetic and acquired causes of pediatric movement disorders; 3 high frequency of MRI lesions 4 that may contribute to patient‐to‐patient differences in functional connectivity; the differing responses observed in distinct monogenic etiologies of dystonia; 5 and difficulty identifying both under‐ and over‐stimulation. Historically, DBS settings have been changed during office visits, guided by patient report and serial neurological examination. Although side effects are usually evident relatively quickly, benefit can be delayed days‐weeks after programming. 6 Dystonia is characteristically variable 7 throughout the day, fluctuating with changes in emotions, physical health, sensory stimulation, and passive/active movement, making it challenging to rely on potentially stressful in‐clinic evaluations, particularly in a nonverbal child.
Recently, DBS systems able to record local field potentials (LFPs) in situ from indwelling DBS electrodes during office visits have become commercially available. LFPs may provide insights that can guide programming, with beta band activity positively correlating with bradykinesia and rigidity in Parkinson's Disease (PD). 8 Both levodopa and DBS stimulation can reduce beta spectral power in PD, correlating with reduced symptom burden. 8 Although promising data has suggested theta power may correlate with severity in adult cervical dystonia, 9 knowledge about potentially clinically‐relevant neurophysiologic biomarkers of dystonia, particularly in the pediatric setting, 10 is incomplete.
We describe a 12 year‐old girl with DYT1‐associated dystonia who underwent bilateral globus pallidus pars interna (Gpi) Medtronic Percept (Dublin, Ireland) SenSight lead placement. At baseline, the patient manifested generalized dystonia affecting the trunk and limbs, worse on the left side of her body (Videos [Link], [Link], [Link]). She experienced chronic daily pain related to her dystonia, struggled to walk, and used a wheelchair for distances. Her baseline Barry‐Albright Dystonia Scale (BADS) score was 17/32, while her Burke‐Fahn‐Marsden (BFM) score was 64/120.
A BrainSense Survey recorded during initial programming demonstrated two prominent LFP peaks, one in the low (13.67 ± 2.5 Hz) and one in the high (25.39 ± 2.5 Hz) beta band (Fig. 1A,B). These peaks appeared at similar frequencies in each hemisphere, although the right GPi peak was more prominent, corresponding to her more severely affected left side.
Figure 1.
(A) Left GPi, initial session using BrainSense Survey (stimulation “Off”) data capture. Vertical lines are linearly arranged from left to right in order of highest average power. Inset depicts the average power for all levels ordered from highest to lowest along the x‐axis. (B) Right GPi, initial session, using same BrainSense Survey (stimulation ‘Off’) data capture. The initial BrainSense Survey suggested optimal levels for stimulation based on LFP data. This was integrated with clinical impression to select levels. The patient was initially discharged from clinic with monopolar DBS settings (left GPi case positive, level 1 negative, current = 1.0 mA, pulse width = 60 μs, frequency 125 Hz; right GPi case positive, level 10 negative, current = 1.5 mA, pulse width = 60 μs, frequency = 125 Hz). (C) Longitudinal changes in local field potentials (LFPs) captured using LFP Timeline data (chronic sensing with stimulation “On”) relative to five distinct programming sessions (April–August) from the right GPi. Each data point reflects a 10‐minute average of power spectral data for a narrow 5 Hz frequency band (low beta: 13.67 ± 2.5 Hz), with distinct colors for each of the five programming visits (four with parameter changes). (D) Zoom‐in depiction of recording period. (E) Left GPi, final session, again using BrainSense Survey (stimulation “Off”) data capture. (F) Right GPi, final session, again using BrainSense Survey (stimulation “Off”) data capture. The patient's final settings were left GPi case positive, level 1 negative, current = 1.0 mA, pulse width = 150 μs, and frequency 125 Hz, while the right GPi settings were case positive, level 10 negative, current = 3.5 mA, pulse width = 180 μs, frequency = 125 Hz. (G) Left GPi first versus last analysis of BrainSense Survey (stimulation “Off”) data using Kolmogorov–Smirnov test (significant values in red) for pairwise contacts. (H) Right GPi first versus last analysis of BrainSense Survey (stimulation “Off”) data using Kolmogorov–Smirnov test statistics (significant values in red) for pairwise contacts. (I) Left GPi analysis of BrainSense Survey (BSS; stimulation “Off”) data between contact pairs. Paired t‐test comparison of low beta vs high beta across contact pairs (t = −3.29(10), P = 0.008) demonstrated a significant difference between first versus last sessions. (J) Right GPi analysis of BrainSense Survey (BSS; stimulation “Off”) data between contact pairs. Paired t‐test analysis of low beta versus high beta across contact pairs (t = −2.13(10), P = 0.059) demonstrated an equivocally significant difference between first vs. last sessions. (K) Change in right GPi LFP power between the 1st and 4th Active Streaming (stimulation “On”) sessions. The first Active Streaming data was captured after the patient's settings had been programmed as reported in (B) for 2 weeks. A two‐way ANOVA (programming session × frequency band) comparing the change in power between sessions by frequency band revealed a significant difference for session for high beta, low beta, alpha and delta (F(1,1158) = 506.0, P < 0.001; individual post‐hoc comparisons exhibited significant differences by band for delta (P < 0.001), theta (P < 0.001), alpha (P < 0.001), low beta (P < 0.001), and gamma (P = 0.01), but not for high beta (P = 0.98)). Although all frequency bands showed a decrease in the final session, high beta appears less responsive to additional stimulation compared to the other frequencies.
Two weeks later, the patient reported her dystonia had improved (Videos S4 and S5) and that her pain had resolved. She now used a mobility scooter instead of a wheelchair for distances. At the third programming session, she noted that a few days after the second visit, her dystonia improved further, but this benefit then waned over the next 10 days, correlating with LFP Timeline data (Fig. 1C,D).
Further details can be found in Supplementary Material (including Figs. [Link], [Link], [Link]). Four months after initial programming, the patient demonstrated only subtle residual dystonia (Video S6) and the patient's mother indicated that she was no longer considered “disabled” by her school. Her BADS score dropped to 3/32 and her BFM score decreased to 3/120 (P < 0.0001; Fisher's exact test) (volume of tissue activation depicted in Fig. S4). Initial to final comparisons with stimulation “off” revealed an appreciable reduction in LFP power, particularly for the high beta peak and more so for the right GPi (Fig. 1E–J), yet changes were observed across the frequency spectrum when comparing early versus late programming sessions with stimulation active (Fig. 1K).
Given the heterogeneity of pediatric dystonia, individualizing programming could help optimize outcomes. We describe an individualized programming approach adopted here based on serial LFP data. Although we acknowledge our index case was a DYT1 patient likely to have a good response to DBS, this allowed us to observe clinical benefit that positively correlated with attenuation of both the high and low beta peaks.
Several important questions remain. It is not known how different electrode placements (ie, GPi vs. subthalamic nucleus) may impact the recorded spectra, nor how widely these peaks (or others) will be observed across dystonia patients of different etiology. It is not known how responsive individual LFPs will be to stimulation across patients. The utility of identifying changes in LFPs during programming sessions to indicate “overstimulation” will also need to be reproduced and extended.
Our data indicate that serial LFP data can be used as a neurophysiologic biomarker to guide clinical decision‐making. Limitations to our study include its intervention‐only nature and single subject design. Although generalization of our findings to a broader group of patients will be key, we anticipate that our data can nevertheless provide key proof‐of‐principle findings supporting the use of LFPs as potentially informative biomarkers for DBS management in dystonia patients.
Author Roles
(1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript: A. Writing of the first draft, B. Review and Critique.
J.K.: 1A, 1B, 1C, 2C, 3A, 3B
A.G.: 1C, 2A, 2B, 3B
P.B.: 1A, 2C, 3B
T.R.L.: 1B, 2B, 2C, 3B
P.B.: 1C, 3B
S.C.: 1C, 3B
P.C.: 1C, 2B, 3B
L.W.: 1C, 3B
F.A.P.: 1B, 1C, 2C, 3B
J.A.T.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B
M.C.K.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B
Disclosures
Ethical Compliance Statement: The study was approved by the Phoenix Children's IRB (#15–080). The patient and her mother provided express written assent/consent for the analysis of her data and publication of findings, including videos. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.
Funding Sources and Conflicts of Interest: No specific funding was received for this work, although JAT and FAP receive research grant support and MCK receives educational grant support from Medtronic, Inc. The authors declare no other real or potential conflicts of interest related to this work.
Financial Disclosures for the Previous 12 Months: JAT's work is supported by NIH UH3 NS113769. MCK's work is supported by NIH R01 NS106298 and R01 NS127108. The authors declare there are no additional disclosures to report.
Supporting information
Data S1. Included are additional descriptions of relevant medical history, surgical placement, and clinical course. Also included in this section are additional details regarding LFP data collection and analysis and the volume of tissue activation analysis along with a summary of key findings.
Figure S1. Stereotactic coordinates. (A) Coordinates are listed as mm values in relation to the midcommissural point and use the (x = lateral, y = anterior, z = superior) convention. The coordinates for the left GPi were (−18. 4.75, 0). (B). Coordinates for the right GPi were (17.75, 4.5, −2). C) Following lead placement, a stereotactic CT was obtained, showing the leads 1.5 and 1.2 mm anterior to the left and right targets, respectively.
Figure S2. Raw visualization. Figure depicts localization of LFP signal across distinct electrode levels (contact sites) and segments.
Figure S3. Within clinic local field potential (LFP) streaming from the right GPi during stimulation amplitude titration (Active Streaming). (A) Titration of stimulation amplitude over the course of the LFP recording. (B–G) Change in LFP power in response to stimulation with isolated representations for delta, theta, alpha, low‐beta, high‐beta and gamma frequency bands. LFP power data were standardized across frequency band by z‐score normalization. In addition, LFP data were smoothed by computing the median for 10 s intervals (original data were sampled at 250 Hz).
Figure S4. DBS electrode reconstruction and volume of tissue activation. (A) Stimulation parameters at the last programming visit. (B, C) Axial and coronal views of the preoperative MRI with overlays of the DBS electrodes computed using postoperative CT. The GPi (green) and calculated volume of tissue activation from stimulation (maroon) are highlighted. (D) Magnified view of each DBS electrode, GPi, and volume of tissue activation.
Video S1. Baseline video examination while seated demonstrates generalized dystonia, worse on the left side of the body, evident at rest and accentuated by certain tasks.
Video S2. Baseline video shows that the patient is able to rise from a seated position and take independent steps, but ambulation is severely impaired by her dystonia.
Video S3. Baseline video reveals the impact of dystonia on gait, with impaired protective responses requiring wall support.
Video S4. Second programming session seated examination shows a reduction in the patient's dystonia.
Video S5. Second programming session stand up and walk evaluation reveals that although dystonia is reduced, the patient is still adjusting to these changes.
Video S6. Final session video captured while walking demonstrates only minimal residual dystonia.
Acknowledgments
We are grateful to the patient and her family for being willing to share her story. We gratefully acknowledge the technical support provided by Bill Valls and Dan Raible PhD (Medtronic) and the programmatic support provided by Nicole Del Col to the Pediatric DBS Program within the Phoenix Children's Barrow Neurological Institute.
References
- 1. Gelineau‐Morel R, Kruer MC, Garris JF, et al. Deep brain stimulation for pediatric dystonia: a review of the literature and suggested programming algorithm. J Child Neurol 2022;37(10–11):813–824. 10.1177/08830738221115248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Fernández‐Alvarez E, Nardocci N. Update on pediatric dystonias: etiology, epidemiology, and management. Degener Neurol Neuromuscul Dis 2012;2:29–41. 10.2147/DNND.S16082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ganaraja VH, Netravathi M, Kamble N, Holla VV, Srinivas D, Yadav R, Pal PK. Spectrum of childhood dystonia evaluated at a tertiary care center from South India. Annal Mov Disord 2021;4(3):136–142. 10.4103/AOMD.AOMD_8_21. [DOI] [Google Scholar]
- 4. Chudy D, Raguž M, Vuletić V, Rački V, Papić E, Nenadić Baranašić N, Barišić N. GPi DBS treatment outcome in children with monogenic dystonia: a case series and review of the literature. Front Neurol 2023;14:1151900. 10.3389/fneur.2023.1151900. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Elkaim LM, Alotaibi NM, Sigal A, et al. Deep brain stimulation for pediatric dystonia: a meta‐analysis with individual participant data. Dev Med Child Neurol 2019;61(1):49–56. 10.1111/dmcn.14063. [DOI] [PubMed] [Google Scholar]
- 6. Macerollo A, Sajin V, Bonello M, Barghava D, Alusi SH, Eldridge PR, Osman‐Farah J. Deep brain stimulation in dystonia: state of art and future directions. J Neurosci Methods 2020;340:108750. 10.1016/j.jneumeth.2020.108750. [DOI] [PubMed] [Google Scholar]
- 7. Aravamuthan BR, Pearson TS, Ueda K, et al. Determinants of gait dystonia severity in cerebral palsy. Dev Med Child Neurol 2023;65(7):968–977. 10.1111/dmcn.15524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Yin Z, Zhu G, Zhao B, et al. Local field potentials in Parkinson's disease: a frequency‐based review. Neurobiol Dis 2021;155:105372. 10.1016/j.nbd.2021.105372. [DOI] [PubMed] [Google Scholar]
- 9. Neumann WJ, Horn A, Ewert S, et al. A localized pallidal physiomarker in cervical dystonia. Ann Neurol 2017;82(6):912–924. 10.1002/ana.25095. [DOI] [PubMed] [Google Scholar]
- 10. Malatt C, Tagliati M. Long‐term outcomes of deep brain stimulation for pediatric dystonia. Pediatr Neurosurg 2022;57(4):225–237. 10.1159/000524577. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1. Included are additional descriptions of relevant medical history, surgical placement, and clinical course. Also included in this section are additional details regarding LFP data collection and analysis and the volume of tissue activation analysis along with a summary of key findings.
Figure S1. Stereotactic coordinates. (A) Coordinates are listed as mm values in relation to the midcommissural point and use the (x = lateral, y = anterior, z = superior) convention. The coordinates for the left GPi were (−18. 4.75, 0). (B). Coordinates for the right GPi were (17.75, 4.5, −2). C) Following lead placement, a stereotactic CT was obtained, showing the leads 1.5 and 1.2 mm anterior to the left and right targets, respectively.
Figure S2. Raw visualization. Figure depicts localization of LFP signal across distinct electrode levels (contact sites) and segments.
Figure S3. Within clinic local field potential (LFP) streaming from the right GPi during stimulation amplitude titration (Active Streaming). (A) Titration of stimulation amplitude over the course of the LFP recording. (B–G) Change in LFP power in response to stimulation with isolated representations for delta, theta, alpha, low‐beta, high‐beta and gamma frequency bands. LFP power data were standardized across frequency band by z‐score normalization. In addition, LFP data were smoothed by computing the median for 10 s intervals (original data were sampled at 250 Hz).
Figure S4. DBS electrode reconstruction and volume of tissue activation. (A) Stimulation parameters at the last programming visit. (B, C) Axial and coronal views of the preoperative MRI with overlays of the DBS electrodes computed using postoperative CT. The GPi (green) and calculated volume of tissue activation from stimulation (maroon) are highlighted. (D) Magnified view of each DBS electrode, GPi, and volume of tissue activation.
Video S1. Baseline video examination while seated demonstrates generalized dystonia, worse on the left side of the body, evident at rest and accentuated by certain tasks.
Video S2. Baseline video shows that the patient is able to rise from a seated position and take independent steps, but ambulation is severely impaired by her dystonia.
Video S3. Baseline video reveals the impact of dystonia on gait, with impaired protective responses requiring wall support.
Video S4. Second programming session seated examination shows a reduction in the patient's dystonia.
Video S5. Second programming session stand up and walk evaluation reveals that although dystonia is reduced, the patient is still adjusting to these changes.
Video S6. Final session video captured while walking demonstrates only minimal residual dystonia.