DBSsync: combining intracranial and multimodal data to investigate new biomarkers in Parkinson’s disease

DBSsync: combining intracranial and multimodal data to investigate new biomarkers in Parkinson’s disease | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article DBSsync: combining intracranial and multimodal data to investigate new biomarkers in Parkinson’s disease Juliette Vivien, Charlotte E. Stensholt, Lucie Hortmann, Merle Hendel, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8228751/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Implanted deep brain stimulation devices are now capable of chronically recording activity from intracranial brain areas during stimulation. This new type of data has the potential to increase our understanding of disease-related brain activity and its modulation in response to therapy or other types of stimuli. With the innovative approach of adaptive deep brain stimulation now clinically available, multimodal characterization of neural biomarkers becomes of utmost importance to define optimal feedback signals for adaptive brain stimulation and allow for better fine-tuning of stimulation parameters. To investigate these biomarkers, we developed DBSsync, a paradigm and an open-source Python toolbox with its graphical user interface for temporally precise synchronization of intracranial recordings with external data, allowing for multimodal research protocols. DBSsync achieves a temporal precision of 8 milliseconds and incorporates cardiac artifact removal methods to facilitate intracranial data preprocessing, thus enabling the integration and precise synchronization of multiple brain signals with external sensors and various behavioral timeline data. Biological sciences/Biological techniques Health sciences/Biomarkers Physical sciences/Engineering Health sciences/Neurology Biological sciences/Neuroscience Figures Figure 1 Figure 2 Figure 3 INTRODUCTION Deep brain stimulation (DBS) is an established and effective treatment for severe movement disorders ( 1 , 2 ). Recently, DBS devices with sensing capacities have become commercially available, widening research horizons by making intracranial electrophysiological recordings accessible to a growing research field. This new feature allows for high-quality local field potential (LFP) recordings in chronically implanted patients, outside of the peri-operative period, therefore with more stable electrophysiology, clinical recovery and optimized therapy. This has led to the discovery of new modulations from known biomarkers such as diurnal modulation of beta power ( 3 ) in patients with Parkinson’s Disease (PD). New biomarkers were also discovered such as stimulation-entrained gamma for motor improvement ( 4 , 5 ) and beta-gamma phase-amplitude coupling related to gait ( 6 ). Discovering personalized neural biomarkers is also important for adaptive DBS (aDBS) as Louie and colleagues have described in five patients treated with aDBS timed to neural biomarkers of contralateral leg swing. Their approach resulted in improved gait compared to continuous DBS treatment ( 7 ). With the rise of these aDBS protocols for better management of both motor and non-motor symptoms in PD ( 8 , 9 ), precise investigation and validation of specific biomarkers become even more essential to decipher complex pathophysiological mechanisms involving multiple brain areas and their link to behavioral readouts. For such a multimodal research approach, the integration and synchronization of multiple brain signals with external sensors and various behavioral timeline data is necessary. Intracranial recordings via sensing-enabled DBS devices are currently transmitted via Bluetooth to a tablet programmed for clinical use. A common connection to a specific recording device that would allow for parallel acquisition of data from additional external sensors such as accelerometers or electroencephalogram (EEG) is therefore not possible. The detection of a shared signal in several recording modalities is needed to perform such synchronization. For example, delivering short DBS pulses can generate artifacts in intracranial recordings that are also detectable near the Percept™ (Percept™ PC/RC, Medtronic, Minneapolis, MN, USA) implantable pulse generator (IPG) ( 10 ). To this date, the reliability and reproducibility of DBS artifacts across patients remain unclear and no open-source and ready-to-use software relying on DBS artifacts for automatic synchronization with multimodal data is yet available. A recent paper ( 11 ) proposed a method and toolbox for the synchronization of DBS data with EEG recordings and used the switch-on artifact of the Percept™ device. However, this toolbox is limited to EEG recordings, not considering data obtained from wearables, videos or other sources that would complete the clinical/behavioral picture during assessment. Additionally, recordings from sensing-enabled deep brain stimulators are often contaminated by electrocardiogram (ECG) artifacts and there is a need for the implementation of reliable and easy-to-use ECG suppression methods ( 12 , 13 ). Here, we describe our open-source Python toolbox called “DBSsync”, offering precise synchronization between chronic intracranial recordings via sensing-enabled DBS devices (Medtronic Percept™) and external recordings of various sources, such as wearables, videos or task-related events, validated in 25 patients with PD and bilaterally implanted with DBS electrodes in the subthalamic nucleus (STN). DBSsync provides an easy-to-use Graphical User Interface (GUI) with multiple features such as ECG artifact cleaning, correction of the Percept™ sampling frequency and verification of proper alignment of the signals after long recording durations. RESULTS Figure 1 shows an example setup and system for the acquisition and synchronization of multimodal data with chronic Percept-LFP data using stimulation pulses and Lab Streaming Layer (LSL). The results presented below are based on three different test datasets, each containing multiple recording sessions (for details see methods section and Table 1 ). Table 1 Overview of participants and session features of each dataset. Dataset Session Time since surgery Digital amplifier sampling frequency Synchronization pulse features STN recording contacts Cardiac artifact in the LFP? 1 sub021 M1S1 24MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1b LSTN: 0–2 N RSTN: 0–2 Y sub024 M0S1 24MFU 4096Hz LSTN: 1mA, 85Hz, 40µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 Y sub033 M1S0 24MFU 4096Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 1–3 Y sub048 M0S1 18MFU 4000Hz LSTN: 1mA, 125Hz, 40µs. Contact 2 LSTN: 1–3 Y RSTN: 0–2 Y sub051 M1S1 18MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 N sub059 M0S0 12MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 1–3 N sub061 M0S1 12MFU 4096Hz LSTN: 1mA, 85Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 1–3 N sub067 M1S0 3MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 N sub070 M1S0 3MFU 512Hz LSTN: 1mA, 145Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 1–3 Y sub084 M1S0 3MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 Y RSTN: 1–3 N sub069 M1S0 12MFU 4096Hz LSTN: 1mA, 125Hz, 40µs. Contact 2 LSTN: 1–3 N 2 RSTN: 1–3 N sub083 M1S1 22MFU 4096Hz LSTN: 1mA, 110Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 Y sub021 M1S1 38MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 1b LSTN: 0–2 N RSTN: 0–2 N sub065 M1S1 17MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 0–2 Y sub065 M1S0 17MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 0–2 Y sub050 M1S1 30MFU 2048Hz LSTN: 1mA, 180Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 1–3 Y sub096 M1S1 10MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 N RSTN: 1–3 N sub052 M1S0 27MFU 2048Hz LSTN: 1mA, 145Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 0–3 N sub047 M1S1 29MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 0–2 Y sub047 M1S0 29MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 0–2 Y sub066 M1S0 21MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 Y sub122 M1S0 7MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 N RSTN: 1–3 N sub078 M1S1 18MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 Y RSTN: 0–2 N sub078 M1S0 18MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 Y RSTN: 0–2 Y sub080 M1S0 17MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 1–2 LSTN: 0–3 N RSTN: 0–2 N sub070 M1S1 21MFU 2048Hz LSTN: 1mA, 145Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 1–3 Y sub093 M1S1 15MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 2 LSTN: 1–3 Y RSTN: 1–3 Y sub043 M1S1 34MFU 2048Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 N sub108 M1S0 10MFU 2048Hz LSTN: 1mA, 85Hz, 60µs. Contact 2 LSTN: 1–3 N RSTN: 1–3 N 3 sub051 M1S0 36MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 N sub051 M1S1 36MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 N sub064 M1S0 24MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 1–3 N sub080 M1S0 24MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 N sub080 M1S1 24MFU 4000Hz LSTN: 1mA, 125Hz, 60µs. Contact 1 LSTN: 0–2 N RSTN: 0–2 N MXSX: medication and stimulation status: 1 for ON, 0 for OFF. MFU (Months Follow-Up): months since deep brain stimulation surgery. LSTN/RSTN: Left and Right Subthalamic Nucleus. Y/N: Yes/No. Recording contacts in the STN are always positioned with one above/one below the stimulation contact(s). Stimulation artifacts are reproducible across patients In order to generate synchronization pulses, two abrupt changes of stimulation amplitude are performed at the beginning and at the end of each recording session by sliding the amplitude up and down quickly (Fig. 1 a, see methods for detailed protocol). These changes in DBS amplitude gave rise to easily identifiable artifacts in both the Percept-LFP data and in an externally recorded bipolar channel placed around the IPG and the cable (Fig. 2 a). In intracranial recordings, the Percept™ hardware suppresses DBS artifacts after a first visible deflection in the signal. In the external bipolar electrode, the artifact is continuously recorded when DBS is turned on and is seen as a fast, repetitive change in signal amplitude, matching the stimulation frequency (Figs. 2 a, 2 b). While the artifact shape in intracranial recordings can be slightly different, it is stable in external recordings even with various sampling frequencies (2048 Hz, 4kHz, 4096Hz). In both signals the polarity of the artifact can vary but this is accounted for automatically in DBSsync (correct detection of polarity: 100% in both signal types). DBSsync enables automatic and reliable synchronization of multimodal data Out of 34 recording sessions, automatic detection of the synchronization artifact was successful in 34 external bipolar channels (100%) and in 33 Percept-LFP channels (97%) (see Fig. 2 b for an example of correct artifact detection). In the remaining Percept-LFP recording, the start of the artifact timepoint was detected two samples too early and was therefore adjusted manually. As described in the methods section, cardiac artifacts were used to determine temporal accuracy of our synchronization method (Fig. 2 c). Using the best sample for synchronization in 8 independent STN (i.e. not recorded during the same session), cardiac artifacts after 10s of recording were never shifted by more than 6ms between Percept-LFP and external ECG, therefore achieving a temporal resolution of 8ms (6ms corresponding to a difference of 1 to 2 samples in the Percept-LFP data sampled at 250Hz). DBSsync helps detect and correct timescale inaccuracies As DBS pulses were repeated at the end of each recording session, the timestamps of the last artifact in the intracranial and external recordings were systematically compared after synchronization to assess data integrity and potential timeshift. As described in the methods section, timeshift may arise after longer recording duration due to small variations in the sampling frequency of the Percept™ device around the 250Hz value indicated and, in less common cases, from the result of packet loss. DBSsync automatically detects packet loss when the .JSON file is directly imported, and adds NaN values in place of the missing data. Out of 34 recording sessions, 3 contained packet loss and were flagged by DBSsync as “corrected for packet loss”. Positive or negative timeshift values were found in all recordings (mean ± std: 11 ± 26ms). For all recording sessions, the effective sampling frequency was computed and varied between 249.985Hz and 250.024Hz (mean ± std: 249.997 ± 0.008Hz), which is a reasonable range given the small variations expected. Therefore, this computed effective sampling frequency was used to create a more accurate timescale for Percept-LFP recordings and synchronize it again with external data, ensuring higher accuracy and consistent alignment of the signals over longer periods of time. After this adjustment of the sampling frequency, timeshift at the end of each recording was computed again and was always close to 0ms (Fig. 2 d). DBSsync enables reliable cleaning of Percept-LFP data from cardiac artifacts 30 out of 68 STN contained visible cardiac artifacts. For each of these recordings, R-peaks were automatically detected, and results were compared with visual verification across 1min of recording (Fig. 3 a). In 2 STN from dataset 1, R-peaks were of small amplitude compared to true brain signal and could not be detected properly because of the absence of an external ECG channel. These 2 STN recordings were therefore not included in subsequent analysis. For each of the 28 remaining STN the best possible method was used (either solely the Percept-LFP data or in combination with a synchronized external ECG channel when available). Across sessions, DBSsync detected overall 97.2% of true positive R-peaks, 0.4% of false positive and 2.4% of R-peaks were missed (Fig. 3 a, see black arrows for missed peaks and grey dots for correctly detected R-peaks). Each of the three cleaning methods was then applied separately and an overlap of raw and cleaned signals was plotted in the GUI (Fig. 3 b shows an example of the result for each method and their associated power spectrums). Power spectrums of raw and cleaned channels were also saved for subsequent analysis and evaluation of each cleaning method. A one-way ANOVA was performed to compare the effect of these three different cleaning methods on ECG suppression ratio. The test revealed that there was a statistically significant difference in ECG suppression ratio between at least two cleaning methods (F (2, 25) = [5.346], p = 0.0067). Post hoc pairwise t -tests with Bonferroni correction indicated that both the template subtraction and SVD method achieved significantly greater ECG suppression compared to the interpolation method (p = 0.0109 and p = 0.0111, respectively). No significant difference was found between the template subtraction and SVD methods. Beta peak recovery was assessed in 7 STN presenting a prominent beta peak. There was no significant difference in beta peak recovery across methods, with all methods achieving values close to 1 indicating good preservation of beta oscillatory activity without spectral enhancement (Fig. 3 d). DISCUSSION In this paper, we present a method to perform temporally precise offline synchronization of intracranial with external recordings using our toolbox DBSsync. One of the major achievements of this toolbox is multimodal data synchronization of at least 5 different signal types (EEG, kinematics, task-related events, physiological recordings (EOG, ECG) and audio) with Percept-LFP data. Following a precise protocol to deliver synchronization pulses, stimulation-induced artifacts were reproducible and reliable, resulting in an accurate automatic detection in most cases. Similar to earlier reports by Soh and colleagues (11), we could demonstrate a temporal precision of 8ms by relying on cardiac artifacts (a method that was more accessible in our setup than relying on Transcranial Magnetic Stimulation (TMS) pulses as they did). In their study, the temporal overlap of TMS pulses was only assessed 10 seconds after synchronization. While this is a good method to validate the accurate synchronization of data at the beginning of a recording, it doesn’t allow us to quantitatively assess the consistency and persistence of time-locking between external and Percept-LFP data across longer recording periods. A verification of the synchronization after 10-20 minutes of recording as proposed in our paradigm is highly valuable, as it is a more realistic duration for behavioral tasks with patients. This verification also revealed an important aspect that had not been addressed in implementations so far: Percept-LFP data can contain timeshift due to either packet loss (short interruptions of connection) or minor variations in the sampling frequency. A strategy to avoid interruptions is to always ensure proximity (<1 meter) and absence of obstacles (even the patient’s own body) between the clinician tablet and the IPG during the recording and avoid the automatic sleep mode by touching regularly the tablet screen. The repetition of synchronization pulses at the end of each recording session allows to easily detect and correct timeshifts by adding NaN values during missing periods to compensate for packet loss and recomputing the effective sampling frequency . Additionally, DBSsync offers a solution to another unmet need in Percept-LFP data preprocessing which is a ready-to-use cardiac artifact removal pipeline. This pipeline demonstrated an accurate and precise R-peak detection, a clear reduction in artifactual low-frequency power achieved mostly by the template subtraction and the SVD methods and a good recovery of beta oscillatory activity. As the SVD method shows similar efficiency results as the template subtraction method but can better adapt to variations in the shape of the R-peak artifact, this method is the one we recommend using, therefore agreeing with the recommendations of Stam and colleagues (13). With respect to the recording possibilities of the Percept TM device, our paradigm can solely be used in the “BrainSense Streaming” mode, as it relies on the induction of stimulation artifacts. Thus, only one bipolar recording per hemisphere is possible with this synchronization technique. To perform synchronization with recordings from all contact pairs (within the so-called “Indefinite Streaming mode” of the Percept PC device), other methods need to be explored. In this regard, a recent paper proposed a different method for synchronization purposes, which relies on computer-driven artifact injection via triggered transcutaneous stimulation (14). This method uses a NeuroOmega neurophysiology system to synchronize intracranial data with task event markers but has not yet been used to synchronize with other external data such as accelerometers or EEG. Our method for synchronization has been developed based on the Percept TM neurostimulator and a TMSi SAGA digital amplifier, but the concept can be adjusted to other neurostimulators with sensing capacity and other data recorders allowing for an external bipolar electrode. Thanks to the versatility of the data types that can be streamed to LSL, our method allows for a high number of possibilities regarding the multimodal acquisition of data due to its compatibility with .XDF data formats. Overall, DBSsync is an open-source toolbox that doesn’t require a large setup or expensive equipment and enables intuitive use through a GUI. It is also versatile in terms of the various recording types that can be synchronized with Percept-LFP data as well as in its preprocessing capacities to improve data quality and temporal precision. Our toolbox therefore provides the opportunity to synchronize intracranial brain activity along with externally recorded signals of interest (cortical EEG, accelerometers, behavioral task events, hand-tracking devices, etc.), for example in the framework of standardized behavioral tasks. This tool will help us realize new research protocols with multimodal recordings integrated to gain further insights into biomarkers, pathophysiology, brain networks and treatment effects in PD. METHODS Patients Twenty-five patients with PD and bilaterally implanted Medtronic SenSight leads in the STN connected to the Percept TM neurostimulator (PC/RC, Medtronic, Minneapolis, MN, USA) were included in the study. Three different dataset types coming from three different studies were used to develop and validate the toolbox (Table 1). Written informed consents were obtained from all patients, and the three studies were approved by the ethics committee at the Charité Universitätsmedizin Berlin (Dataset 1: EA2_256_20, Datasets 2 and 3: EA1/164/23) and conducted following the standards set by the Declaration of Helsinki. Datasets Dataset 1 including accelerometer data and LFP : recordings were performed 3-24 months after DBS surgery as part of a long-term study investigating finger-tapping movements using accelerometers and intracranial LFP both with and without active DBS (n = 10 recording sessions). Dataset 2 including EEG and LFP during a behavioral task : recordings were performed 7-38 months after DBS surgery as part of a study recording cortical and intracranial LFP during cued button presses both with and without active DBS (n = 19 recording sessions). Dataset 3 including hand movements in 3D-space, audio and LFP during a behavioral task : recordings were performed 24-36 months after DBS surgery as part of a study recording kinematics, audio and intracranial LFP during a behavioral task both with and without active DBS (n = 5 recording sessions). Recording setup For all datasets, intracranial brain activity was streamed from bilaterally implanted DBS-electrodes with a sampling frequency of 250 Hz via a Bluetooth connection to a tablet used for clinical programming. As external recording source for the synchronization artifact, a bipolar electrode with one contact placed close to the IPG and the other placed close to the IPG cable was connected to a digital amplifier (TMSi International, Oldenzaal, NL – TMSi SAGA) (Figure 1a). In dataset 1, two tri-axial accelerometers placed on both index fingers were recorded via the same digital amplifier. The output file of our digital amplifier was a .POLY5 file. In dataset 2, the digital amplifier was also connected to a 32-channels EEG cap (BrainWave Cap Infinity – 32ch+gnd, Ag/AgCl – electrodes), an accelerometer placed on the finger used to press a response button during a behavioral task and two more bipolar electrodes used as electrooculogram (EOG) and ECG channels. Task events and signals from the digital amplifier were sent as two separate streams to the Lab Recorder application using the Lab Streaming Layer framework (LSL, 15) which allowed for online automatic synchronization of these data streams. Using the Lab Recorder application, each recording resulted in a single output file in the eXtensible Data Format (.XDF) containing all synchronized external streams (Figure 1b). In dataset 3, the digital amplifier was solely used for recording the bipolar electrode used for synchronization (around the IPG). LSL was also used during the experiment and received four different streams: one from the digital amplifier containing the signal from the bipolar electrode, one from the behavioral task containing the event markers, one from the Ultraleap camera (Leap Motion Controller 2 (Ultraleap Ltd., Bristol, UK, 2023) or Stereo IR 170 Camera (Ultraleap Ltd., Bristol, UK, 2020) recording hand movements in 3D space and one from the audio recorder. These four streams resulted in a single .XDF file as output. The output file of Percept-LFP recordings was a JavaScript Object Notation (.JSON) file, containing all recordings and metadata from an experimental session. Induction of DBS synchronization artifacts We established a specific paradigm to induce stimulation artifacts in intracranial and external recordings for later data alignment. This paradigm consisted of two stimulation pulses at the beginning and end of each recording session (purple line in Figure 1a). Specifically, the ramp option was deactivated to allow for large amplitude pulses and baseline recordings were started with DBS turned on and bilaterally set to 0mA for 5s. Two short pulses (approximately 2s with an interval of 2s each) of unilateral, high-frequency DBS at 1mA amplitude were then delivered. DBS was then either kept at 0mA (DBS-Off sessions, n=17) or manually ramped up to the amplitude(s) tested in the study (DBS-On sessions, n=17) (Table 1). Before ending each recording, DBS was set back to 0mA bilaterally and two sequences of unilateral, high-frequency DBS with the same settings and in the same hemisphere as in the beginning were delivered. Main interface and compatible file formats DBSsync is a Python-based open-source toolbox developed to align intracranial Percept-LFP recordings and external recordings (available at: https://github.com/juliettevivien/DBSsync.git). The toolbox is designed with a GUI for easier use, and a user guide is available in the supplementary data. Within DBSsync, a Percept-LFP recording from one session (.JSON, .MAT or .FIF file) and the corresponding external recording from the same session (.XDF, .POLY5 or .FIF file) are loaded. If the external file is .XDF, the user is first asked to select which LSL stream contains the bipolar electrode used for synchronization. MAT file compatibility was implemented and designed for users who use the open-source Perceive toolbox (available at https://github.com/neuromodulation/perceive/) to extract each recording session as a single .MAT file. FIF files compatibility was added for both external and intracranial files to increase compatibility for MNE-python users (16; MNE-python) who might have their own preprocessing pipeline for .JSON files and/or different output formats for the external recordings. Synchronization of intracranial Percept-LFP and external recordings In the time series of the external bipolar electrode recorded with the digital amplifier, the DBS artifact is detected as a steep and sustained decrease/increase (depending on electrode polarity) in signal amplitude after applying a high pass filter (1Hz) to detrend the data. The sample with the first highest amplitude change from baseline is selected as “start of the artifact” in the external channel (grey dashed line in Figure 2b upper plot). In the intracranial recording, the DBS artifact is a sharp change in signal amplitude of the signal followed by a slow recovery (Figure 2a). To identify at which point the sharp change occurs, a threshold window is computed based on the first two seconds of each intracranial recording. The last sample that lies within the value distribution of the window before crossing the threshold is detected as the start of the artifactual period (sample 0 in Figure 2c). As the artifact in the Percept-LFP data spreads over several samples and doesn’t have a consistent shape across patients, we used a numerical labeling of each sample after this sharp change. To test which sample should be chosen as actual “start of the artifact”, 8 STN from independent sessions containing clear cardiac artifacts in both the Percept-LFP data and in an external ECG channel were used. Each sample was successively used for synchronization and the alignment of endogenous cardiac artifacts R-peaks in intracranial and external recordings was assessed 10 seconds later to reproduce the analysis performed with TMS pulses by Soh and colleagues (11) for comparison. In all artifacts’ shape, the sample who overall provided the best temporal alignment of endogenous cardiac artifacts was the 4 th sample after the sharp change (black sample in Figure 2c) (mean time difference across 8 recording sessions ± std: 0 ± 4ms, min: -6ms, max +6ms). Therefore, the 4 th sample is defined as “start of the artifact” and is always used for synchronization in DBSsync (grey dashed line in Figure 2b lower plot). In the GUI, a plot shows the sample automatically chosen as the start of the artifact in both time series, allowing the user to decide whether the artifact has been properly detected. If the automatic detection method fails due to an unusual artifact shape or a baseline contaminated by other artifacts, the user can correct the DBS artifact detection and manually select the correct sample timepoint. Verification of the consistency of synchronization over time In the timeshift window, the two synchronized channels can be plotted and overlapped to check the consistency of the synchronization over using artifacts generated at the end of the recording sessions (Figure 2d). DBSsync offers the possibility to compute the timeshift (i.e. the time-delay between these end-artifacts, see bottom plots in Figure 2d). If the absolute value of the timeshift is high (>200ms), it suggests that data loss happened in the Percept-LFP data, most times due to communication issues between the IPG and the clinician tablet. Such recordings should not be analyzed further before properly handling this data loss. If the Percept-LFP data wasn’t loaded using the original .JSON file, consider using this method as it automatically corrects for packet loss. If the absolute value of the timeshift is below this value but still larger than 0, it might reflect a slight variation in the sampling frequency of the Percept-LFP data from 250Hz. This might not be considered relevant during short experiments but for longer recording sessions it produces inconsistent synchronization of the signals towards the end (up to 65 milliseconds of timeshift was found in our datasets for a recording of ~11minutes). Therefore, DBSsync also encompasses the possibility to compute the effective sampling frequency of the Percept-LFP data using the sampling frequency of the external digital amplifier as ground truth (usually much higher sampling frequency and data are not sent via Bluetooth, but via USB connection which allows for higher accuracy in sampling frequency information). Once this effective sampling frequency is computed, it is automatically added to the Percept-LFP metadata and the synchronization artifact must be detected again in the LFP channel using this new sampling frequency for better accuracy. When computing the timeshift again, this results in a value close to 0ms (Figure 2d, right lower plot). Cardiac artifact removal from Percept-LFP data DBSsync offers the possibility to clean Percept-LFP data from cardiac artifacts either after synchronization with external data or independently. The first step is to detect the R-peaks in the LFP channel (Figure 3a). The contaminated channel can be used on its own or in combination with a synchronized external ECG channel recorded through the digital amplifier (recommended for better detection of R-peaks). A low-pass filter can be applied to the LFP data prior to cleaning, to remove potential stimulation artifacts and enhance signal-to-noise ratio. If the Percept-LFP data is used alone, the detection algorithm will segment the signal into overlapping 1-second windows and search for negative and positive peaks in each window to get a first approximation of the R-peaks localization and their polarity. 1-second epochs centered on these R-peaks are generated and averaged to create a first ECG template. The cross-correlation between this ECG template and the signal is calculated and R-peaks are identified if they exceed a 95 th percentile threshold with a minimum distance of 0.5s between peaks. A refined ECG template is computed, and a second-pass detection is performed in which the threshold can be modified by the user to refine the detection. If an external ECG channel is synchronized and available to help detect R-peaks, this channel is band-pass filtered between 0.5Hz and 60Hz to remove slow drifts and potential DBS artifacts and the algorithm searches for the timestamps of R-peaks in the ECG using a 95th percentile threshold and a minimum distance of 0.5s between peaks. Corresponding R-peaks are searched in the LFP data by creating time windows from -80ms to +80ms around each R-peak timepoint as detected in the ECG (the algorithm searches for local minima/maxima values in each window in the LFP data). These two detection methods are reproduced from the description provided by Stam and colleagues (13). DBSsync offers the possibility to manually override some of the parameters for better control of R-peak detection (e.g. wrong polarity detected, artifactual periods to avoid, percentile threshold to use). The second step is to clean the signal from ECG artifacts based on the detected R-peaks. The methods implemented in DBSsync are also adapted from the paper published by Stam and colleagues (13), in which they extensively describe three different pipelines for cardiac artifact removal in Percept-LFP recordings. These three methods are available in DBSsync to dampen the cardiac artifact: 1) an interpolation method, in which the R-peaks are replaced by linearly interpolated data, 2) a template subtraction method, in which an average template of the QRS complex is generated, which is then fitted and subtracted to each cardiac artifact individually, and 3) a Singular Value Decomposition (SVD) method, in which the cardiac artifact is reconstructed at each R-peak based on the first “k” components of the SVD, fitted and subtracted separately, to account for differences in artifact shape. The number “k” of SVD components to use for the reconstruction is set by the user, based on the shape (as described in 13) and explained variance of each component. DBSsync automatically displays four output plots to assess the quality of cardiac artifact removal: 1) detected R-peaks (Figure 3a), 2) ECG artifact shape (average template across epochs or average reconstruction using SVD components), 3) an overlap of the raw and cleaned channel (Figure 3b, left) and 4) an overlap of the power spectrum of the raw and cleaned data (Figure 3b, right). Once cardiac artifacts are properly removed, the user can choose to replace the original channel with the cleaned one. The original channels will still be saved in the output file but labeled as “RAW”. Validation of cardiac artifact removal Across 68 STN recorded, 30 contained visible cardiac artifacts (Table 1). R-peak detection was performed in these STN recordings without using any external ECG channel and the accuracy of the detection was assessed by counting true positive, false positive and missed peaks over one minute of recording (Figure 3a). The minute chosen for testing R-peak detection accuracy was always from 80 to 140s, as it was the first common minute free of synchronization pulses in all recordings. When available, this was performed again using an external ECG channel to help R-peaks detection. The best detection was kept for further cleaning of the channel. In some cases, the "manual override" option was used to correct for ECG artifact polarity (5 cases) or the start/end of detection (to avoid stimulation pulses, which can bias the ECG artifact shape if included in the template). Power spectrums of raw and cleaned channels were computed using Welch’s method with a 1s window and 50% overlap. These power spectrums were used to determine ECG suppression ratio and beta peak recovery achieved by each method. These metrics were chosen and adapted from the thesis of Silvi L. (17). The ECG suppression ratio measures power attenuation with higher values indicating greater cardiac artifact removal. It is calculated as: The peak prominence is calculated relative to the surrounding spectral baseline within an extended frequency window (±5 Hz around peak frequency). In our datasets, 7 STN contained a prominent beta peak and were used for this analysis. For statistical assessment of each metric, a one-way ANOVA was performed to evaluate overall differences between the three cleaning methods, followed by Bonferroni-corrected post-hoc pairwise t -tests to identify specific significant differences between methods. Saving synchronized and/or preprocessed recordings The saving option can be set by the user in the config file before starting DBSsync. Available saving options are .SET, .FIF, .MAT or .PKL. In the case of external .XDF files, users should be cautious of what type of stream was contained in it: if it only contained continuous streams and marker streams, the synchronized/cleaned data can be saved in .SET, .MAT or .FIF format. However, if the .XDF file contained discontinuous streams (i.e. without a defined sampling frequency but only discrete data points) the synchronized data must be saved as .PKL (as the other formats only allow for one common sampling frequency for all channels). Declarations DATA AVAILABILITY Data are available conditionally through data-sharing agreements in accordance with data privacy statements signed by the patients within the legal framework of the General Data Protection Regulation of the European Union within a time frame of 6 months. Requests should be directed to A.A.K. ( [email protected] ) or the Open Data officer ( [email protected] ). CODE AVAILABILITY The code of the toolbox is publicly available in a Github repository and can be accessed via this link: https://github.com/juliettevivien/DBSsync ACKNOWLEDGEMENTS We would like to thank our patients for their participation in our research protocols. We would also like to thank Jennifer K. Behnke, Varvara Mathiopoulou, Johannes L. Busch and Aicha Ben Jannet who contributed in part to data collection from dataset 1. FUNDING DECLARATION JV has received a Doctoral Research Grant from the German Academic Exchange Service – Deutscher Akademischer Austauschdienst (DAAD). RL and JGVH are participants in the BIH Charité Clinician Scientist Program funded by the Charité – Universitätsmedizin Berlin, and the Berlin Institute of Health at Charité (BIH). This work was supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Project-ID 424778381 – TRR 295; and by the Lundbeck Foundation as part of the collaborative project grant ADAPT-PD (Grant Nr. R336-2020-1035). AUTHOR CONTRIBUTIONS JV has designed the methodology, performed experiments, collected and analyzed data, developed the software and graphical user interface, wrote original drafts and made revisions, created scientific figures and acquired funding. CES, LH, MH, AM and LKF have performed experiments and collected data. CES also participated in data analysis, tested code components and reviewed and edited the original draft of the manuscript. LKF and RL were involved in conceptualization and provided critical review and editing to the manuscript. JGVH and AC havecontributed to the programming and testing of existing code components. AAK was involved in conceptualization, critical review and editing of the manuscript, supervision and provided resources and funding. All authors read and approved of the final manuscript. COMPETING INTERESTS A.A.K. has served on advisory boards of Medtronic and has received honoraria and travel support from Medtronic, Boston Scientific, Ipsen Pharma and Teva. L.K.F. has received speaker honoraria from Medtronic. References Deuschl, G., et al. A randomized trial of deep-brain stimulation for Parkinson's disease. The New England journal of medicine, 355(9), 896–908 (2006). Volkmann, J. et al. Pallidal neurostimulation in patients with medication-refractory cervical dystonia: a randomised, sham-controlled trial. The Lancet Neurology 13, 875–884 (2014). Van Rheede, J. J. et al. Diurnal modulation of subthalamic beta oscillatory power in Parkinson’s disease patients during deep brain stimulation. npj Parkinsons Dis. 8, 88 (2022). Mathiopoulou, V. et al. Gamma entrainment induced by deep brain stimulation as a biomarker for motor improvement with neuromodulation. Nat Commun 16, 2956 (2025). Olaru, M. et al. Deep brain stimulation-entrained gamma oscillations in chronic home recordings in Parkinson’s disease. Brain Stimulation 18, 132–141 (2025). Farokhniaee, A. et al. Gait-related beta-gamma phase amplitude coupling in the subthalamic nucleus of parkinsonian patients. Sci Rep 14, 6674 (2024). Louie, K. H., Balakid, J. P., Azgomi, H. F., Marks, J. H. & Wang, D. D. Adaptive deep brain stimulation timed to gait phase improves walking in Parkinson’s disease. medRxiv : the preprint server for health sciences (2025) Busch, J. L. & Kaplan, J. et al. Chronic adaptive deep brain stimulation for Parkinson’s disease: clinical outcomes and programming strategies. npj Parkinsons Dis. 11, 264 (2025). Swinnen, B. E. K. S., et al. Basal ganglia theta power indexes trait anxiety in people with Parkinson's disease. Brain: a journal of neurology, 148(4), 1228–1241 (2025). Thenaisie, Y. et al. Towards adaptive deep brain stimulation: clinical and technical notes on a novel commercial device for chronic brain sensing. J. Neural Eng. 18, 042002 (2021). Soh, C., Hervault, M., Rohl, A. H., Greenlee, J. D. W. & Wessel, J. R. Precisely-timed outpatient recordings of subcortical local field potentials from wireless streaming-capable deep-brain stimulators: a method and toolbox. Journal of Neuroscience Methods 418, 110448 (2025). Hammer, L. H., Kochanski, R. B., Starr, P. A. & Little, S. Artifact Characterization and a Multipurpose Template-Based Offline Removal Solution for a Sensing-Enabled Deep Brain Stimulation Device. Stereotact Funct Neurosurg 100, 168–183 (2022). Stam, M. J. et al. A comparison of methods to suppress electrocardiographic artifacts in local field potential recordings. Clinical Neurophysiology 146, 147–161 (2023). Alarie, M. E., Provenza, N. R., Herron, J. A. & Asaad, W. F. Automated artifact injection into sensing-capable brain modulation devices for neural-behavioral synchronization and the influence of device state. Brain Stimulation 16, 1358–1360 (2023). Kothe, C. et al. The lab streaming layer for synchronized multimodal recording. Imaging Neuroscience 3, IMAG.a.136 (2025). Gramfort, A. MEG and EEG data analysis with MNE-Python. Front. Neurosci. 7, (2013). Silvi, L. Electroencephalography and Local Field Potential fusion to characterize Deep Brain Stimulation in Parkinson’s disease. [Master’s thesis] (2025) Additional Declarations Competing interest reported. AAK has served on advisory boards of Medtronic and has received honoraria and travel support from Medtronic, Boston Scientific, Ipsen Pharma and Teva. LKF has received speaker honoraria from Medtronic. Supplementary Files supplementarydataDocumentationDBSsyncGUI.pdf Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 01 Mar, 2026 Reviews received at journal 26 Feb, 2026 Reviews received at journal 16 Jan, 2026 Reviewers agreed at journal 05 Jan, 2026 Reviewers agreed at journal 05 Jan, 2026 Reviewers agreed at journal 05 Jan, 2026 Reviewers invited by journal 04 Dec, 2025 Editor assigned by journal 02 Dec, 2025 Submission checks completed at journal 02 Dec, 2025 First submitted to journal 28 Nov, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8228751","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":556299737,"identity":"99402557-a1a1-4afc-9937-b101da06cc03","order_by":0,"name":"Juliette Vivien","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Juliette","middleName":"","lastName":"Vivien","suffix":""},{"id":556299738,"identity":"980aaf72-12cd-4ed6-80c2-156d3c91f33d","order_by":1,"name":"Charlotte E. Stensholt","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Charlotte","middleName":"E.","lastName":"Stensholt","suffix":""},{"id":556299739,"identity":"f263c7fb-4720-4909-8b3b-2272731311f2","order_by":2,"name":"Lucie Hortmann","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Lucie","middleName":"","lastName":"Hortmann","suffix":""},{"id":556299740,"identity":"33eca467-23dd-4ad8-b826-4b33671ae854","order_by":3,"name":"Merle Hendel","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Merle","middleName":"","lastName":"Hendel","suffix":""},{"id":556299743,"identity":"645697aa-9422-4db0-895f-85461532c9cd","order_by":4,"name":"Arian Memarpouri","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Arian","middleName":"","lastName":"Memarpouri","suffix":""},{"id":556299748,"identity":"5da0c7d2-b577-4c4d-8f50-22986af79670","order_by":5,"name":"Roxanne Lofredi","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Roxanne","middleName":"","lastName":"Lofredi","suffix":""},{"id":556299749,"identity":"3cd4d629-7781-4609-a417-1138c61594cb","order_by":6,"name":"Jeroen G. V. Habets","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Jeroen","middleName":"G. V.","lastName":"Habets","suffix":""},{"id":556299750,"identity":"0256d14f-19ad-4ab0-b27b-6020a88cc4a6","order_by":7,"name":"Alessia Cavallo","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Alessia","middleName":"","lastName":"Cavallo","suffix":""},{"id":556299751,"identity":"13297e42-cbe1-435f-88d2-14ec33ae3741","order_by":8,"name":"Lucia K. Feldmann","email":"","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":false,"prefix":"","firstName":"Lucia","middleName":"K.","lastName":"Feldmann","suffix":""},{"id":556299753,"identity":"51a7a91f-c27d-4f17-9b0b-e229609c8270","order_by":9,"name":"Andrea A. Kühn","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAklEQVRIiWNgGAWjYHACAyjBw8DwgeEAAwMziM9GpBbGGWAtzCRoYeYBaWEgoIV/dvO2Dx9qGIwNjvcee2xTc0fevJ3/4AOGMhucWiTuHCueOeMYg5nBmXPpxjnHnhnOOczMbMBwLg23NTdyjJl52BhsDG7kmEnnNhxmnMHMzCbB2HYYpw55kJY//4Ba7r8xk7ZsOGwP1ML+g7HtP26/g7QwtgEddoPHTJqx4XAiyBYGxrYDOLUY3kgrZuztkzCWPJNjJtlz7HAyUIuxRMK5ZJxa5G4kb2b48c3GsO/4GTOJHzWHbWfwH3z44UOZHW7vQ4AEGj+BkIZRMApGwSgYBXgBAMCsT5C22D5oAAAAAElFTkSuQmCC","orcid":"","institution":"Charité-Universitätsmedizin Berlin","correspondingAuthor":true,"prefix":"","firstName":"Andrea","middleName":"A.","lastName":"Kühn","suffix":""}],"badges":[],"createdAt":"2025-11-28 09:23:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8228751/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8228751/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97699197,"identity":"f84246d6-6b9d-4c8c-aeb0-7273e2ca584f","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":566106,"visible":true,"origin":"","legend":"","description":"","filename":"DBSsynccombiningintracranialandmultimodaldatatoinvestigatenewbiomarkersinParkinsonsdisease.docx","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/660474374f9e56b2a6a7c334.docx"},{"id":97699195,"identity":"e0d7e7ca-ff3d-48de-97d8-e153fdfb1bb9","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11300,"visible":true,"origin":"","legend":"","description":"","filename":"a3a862502fea48838aae0df2ebf59843.json","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/aa10047809aeaddb111caef1.json"},{"id":97699208,"identity":"44354975-b421-479c-a64d-142952847253","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2393294,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarydataDocumentationDBSsyncGUI.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/a7232ed133d46bff693e6b19.pdf"},{"id":97699203,"identity":"351ede6f-f470-4f30-8f3e-726ddea9245c","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":103886,"visible":true,"origin":"","legend":"","description":"","filename":"a3a862502fea48838aae0df2ebf598431enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/2c40e49519a10d9ec4c0ca8c.xml"},{"id":97699194,"identity":"46c88749-e32c-4c69-826e-f98abe651b6e","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":134875,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/0fbee71f7d3910ea3cf97e99.png"},{"id":97699196,"identity":"b86b0adc-c976-44e3-b107-bd6634848d1e","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":192683,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/a6142661bd527569525d7a0c.png"},{"id":97894374,"identity":"bb3d3f3a-335f-4088-bfb9-1d6c5e38f37c","added_by":"auto","created_at":"2025-12-10 15:32:25","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":208394,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/910dca4d19736d41cdeea212.png"},{"id":97895044,"identity":"b42e404c-cd10-4079-9de5-ba6981634805","added_by":"auto","created_at":"2025-12-10 15:33:26","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39494,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/2680b99bcdd0003ce4af73f1.png"},{"id":97699199,"identity":"3e105887-1424-4340-92cb-f9a3f3bd036a","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":74166,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/0560da391d2780abdd5d09bb.png"},{"id":97699206,"identity":"c09a4814-a1e0-45ca-b306-f1bf02d051bc","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":75991,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/74d30776a6f8dfc85fb5157e.png"},{"id":97699205,"identity":"bee838e9-a954-4550-8fb3-29e70662366b","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102636,"visible":true,"origin":"","legend":"","description":"","filename":"a3a862502fea48838aae0df2ebf598431structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/886e3e33f062595893d9dd2d.xml"},{"id":97699202,"identity":"ee4f1a4f-bcb2-4cb2-9952-36e72191a1de","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"html","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111875,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/15cee07cdfc9b8e5b2d80460.html"},{"id":97894137,"identity":"b48a86df-ce3e-4d31-b3e6-61eac55f83b8","added_by":"auto","created_at":"2025-12-10 15:31:58","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":168974,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRecording setup.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea.\u003c/strong\u003e Example of setup for dataset 2: the patient is equipped with three bipolar electrodes (“BIPs”): one is used for detecting DBS synchronization pulses and is placed close to the IPG and cable, another is recording eye movements (EOG channel) and the last is recording the heartbeat (ECG channel). An accelerometer (ACC) is placed on the index finger to record button presses during a behavioral task, and an EEG cap with 32 channels is also used to record cortical electrophysiological activity. All these channels are recorded via an external digital amplifier, which is streamed to the Lab Recorder Application. The activity from the STN is recorded from implanted DBS electrodes and sent via Bluetooth connection to the clinician recording tablet. Two synchronization DBS pulses are performed at the beginning and at the end of each recording session. GND: ground.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb.\u003c/strong\u003e The Lab Recorder application can receive streams from different sources (external digital amplifier, behavioral task, video camera, etc.). It automatically synchronizes all streams and, at the end of the recording session, saves one single output file in .XDF format, with a common timeline.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/8895304d830f00b40d88f3c9.png"},{"id":97894165,"identity":"9badd740-a9d9-4565-9d98-019fc32b4ca6","added_by":"auto","created_at":"2025-12-10 15:31:59","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":232709,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSynchronization of intracranial and external recordings.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea.\u003c/strong\u003e In the external bipolar electrode, a DBS synchronization pulse induces a fast and repetitive deflection in the amplitude (top plot), whereas in the intracranial channel it induces a sharp drop in the amplitude of the signal, followed by a slow recovery (lower plot).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb.\u003c/strong\u003e Zoom on panel \u003cstrong\u003ea\u003c/strong\u003e, grey dashed lines show which sample is chosen as start of the artifact in each recording modality for the synchronization in DBSsync.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec.\u003c/strong\u003e Method used to investigate which sample of the DBS artifact to choose as “start of the artifact” in the Percept-LFP timeseries. The blue dotted line represents a typical artifact shape as seen in the intracranial channel from panel \u003cstrong\u003eb\u003c/strong\u003e. The last sample before the sharp change in amplitude is labeled as sample 0. Different samples were chosen for the synchronization to test which sample provides the best alignment of endogenous cardiac artifacts between the intracranial channel and an externally acquired ECG channel. As depicted in this example, sample number 4 provided the best results, here with a delay of -1ms observed between the two cardiac artifacts. This sample is always chosen for synchronization in the toolbox, as depicted in panel \u003cstrong\u003eb\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed.\u003c/strong\u003e Full recording session with external bipolar electrode (top) and intracranial (bottom) channels synchronized. The two synchronization pulses at the end (dashed box) are used to calculate the timeshift between the two recordings. After more than 1400s of recording, as depicted on the lower plots, the last synchronization pulse is visible with a delay of 46ms in the two synchronized channels (left). After a correction of the sampling frequency (sf) of the Percept-LFP data, the synchronization pulse is now perfectly aligned at the end of the two channels (right).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/854aa7fbe3e382a2096c7058.png"},{"id":97699192,"identity":"1f9aeb95-d330-4e31-8858-52314294c70a","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":234519,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eECG artifact cleaning in Percept-LFP data.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ea.\u003c/strong\u003e R-peak detection and visual verification. R-peaks automatically detected in the Percept-LFP channel (in blue) by DBSsync are labeled with grey dots. In this example, two R-peaks were missed (identified with black arrows). This visual verification over 1 minute of recording was performed in all analyzed sessions and overall DBSsync performance was calculated and reported in the confusion matrix on the right side.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eb.\u003c/strong\u003e ECG artifact removal using three different methods. Raw Percept-LFP time series (dark line) overlapped with cleaned data after using linear interpolation (top), template subtraction (middle) or Singular Value Decomposition (bottom) to remove ECG artifact. The resulting power spectrum for each time series is shown on the right side.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ec.\u003c/strong\u003e ECG suppression ratio (in the 0.5-40Hz frequency band) achieved by all three cleaning methods (n = 28 STN, mean ± SD). Template subtraction and SVD methods show clear superiority over the linear interpolation method (p = 0.0109 and 0.0111, respectively).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ed.\u003c/strong\u003e Beta peak recovery achieved by all three cleaning methods. Only prominent peaks in the 13-35Hz frequency range were used (n = 7 STN, mean ± SD). There was no significant differences between methods.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/81b0d09f786d1a08e5da4493.png"},{"id":98421025,"identity":"9dcbe1cc-d07f-48b8-9c32-d7d627d192cb","added_by":"auto","created_at":"2025-12-17 16:22:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1729351,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/35be3895-ac5b-4400-a87b-5d0b9bd338e4.pdf"},{"id":97699204,"identity":"38232ba6-c0b0-48c6-b0ab-88a466e0a524","added_by":"auto","created_at":"2025-12-08 12:03:22","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":2393294,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarydataDocumentationDBSsyncGUI.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8228751/v1/f56473208982a01fd01f463e.pdf"}],"financialInterests":"Competing interest reported. AAK has served on advisory boards of Medtronic and has received honoraria and travel support from Medtronic, Boston Scientific, Ipsen Pharma and Teva.\nLKF has received speaker honoraria from Medtronic.","formattedTitle":"\u003cp\u003eDBSsync: combining intracranial and multimodal data to investigate new biomarkers in Parkinson’s disease\u003c/p\u003e","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eDeep brain stimulation (DBS) is an established and effective treatment for severe movement disorders (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). Recently, DBS devices with sensing capacities have become commercially available, widening research horizons by making intracranial electrophysiological recordings accessible to a growing research field. This new feature allows for high-quality local field potential (LFP) recordings in chronically implanted patients, outside of the peri-operative period, therefore with more stable electrophysiology, clinical recovery and optimized therapy. This has led to the discovery of new modulations from known biomarkers such as diurnal modulation of beta power (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) in patients with Parkinson\u0026rsquo;s Disease (PD). New biomarkers were also discovered such as stimulation-entrained gamma for motor improvement (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e) and beta-gamma phase-amplitude coupling related to gait (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). Discovering personalized neural biomarkers is also important for adaptive DBS (aDBS) as Louie and colleagues have described in five patients treated with aDBS timed to neural biomarkers of contralateral leg swing. Their approach resulted in improved gait compared to continuous DBS treatment (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). With the rise of these aDBS protocols for better management of both motor and non-motor symptoms in PD (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e), precise investigation and validation of specific biomarkers become even more essential to decipher complex pathophysiological mechanisms involving multiple brain areas and their link to behavioral readouts. For such a multimodal research approach, the integration and synchronization of multiple brain signals with external sensors and various behavioral timeline data is necessary.\u003c/p\u003e\u003cp\u003eIntracranial recordings via sensing-enabled DBS devices are currently transmitted via Bluetooth to a tablet programmed for clinical use. A common connection to a specific recording device that would allow for parallel acquisition of data from additional external sensors such as accelerometers or electroencephalogram (EEG) is therefore not possible. The detection of a shared signal in several recording modalities is needed to perform such synchronization. For example, delivering short DBS pulses can generate artifacts in intracranial recordings that are also detectable near the Percept\u0026trade; (Percept\u0026trade; PC/RC, Medtronic, Minneapolis, MN, USA) implantable pulse generator (IPG) (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). To this date, the reliability and reproducibility of DBS artifacts across patients remain unclear and no open-source and ready-to-use software relying on DBS artifacts for automatic synchronization with multimodal data is yet available. A recent paper (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e) proposed a method and toolbox for the synchronization of DBS data with EEG recordings and used the switch-on artifact of the Percept\u0026trade; device. However, this toolbox is limited to EEG recordings, not considering data obtained from wearables, videos or other sources that would complete the clinical/behavioral picture during assessment. Additionally, recordings from sensing-enabled deep brain stimulators are often contaminated by electrocardiogram (ECG) artifacts and there is a need for the implementation of reliable and easy-to-use ECG suppression methods (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHere, we describe our open-source Python toolbox called \u0026ldquo;DBSsync\u0026rdquo;, offering precise synchronization between chronic intracranial recordings via sensing-enabled DBS devices (Medtronic Percept\u0026trade;) and external recordings of various sources, such as wearables, videos or task-related events, validated in 25 patients with PD and bilaterally implanted with DBS electrodes in the subthalamic nucleus (STN). DBSsync provides an easy-to-use Graphical User Interface (GUI) with multiple features such as ECG artifact cleaning, correction of the Percept\u0026trade; sampling frequency and verification of proper alignment of the signals after long recording durations.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows an example setup and system for the acquisition and synchronization of multimodal data with chronic Percept-LFP data using stimulation pulses and Lab Streaming Layer (LSL). The results presented below are based on three different test datasets, each containing multiple recording sessions (for details see methods section and Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eOverview of participants and session features of each dataset.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDataset\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSession\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTime since surgery\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eDigital amplifier sampling frequency\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eSynchronization pulse features\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eSTN recording contacts\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eCardiac artifact in the LFP?\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"19\" rowspan=\"20\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub021 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e24MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1b\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub024 M0S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e24MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4096Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 85Hz, 40\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub033 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e24MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4096Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub048 M0S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e18MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 40\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub051 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e18MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub059 M0S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e12MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub061 M0S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e12MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4096Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 85Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub067 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e3MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub070 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e3MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e512Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 145Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub084 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e3MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub069 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e12MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4096Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 40\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"36\" rowspan=\"37\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub083 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e22MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4096Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 110Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub021 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e38MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1b\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub065 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e17MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub065 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e17MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub050 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e30MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 180Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub096 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e10MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub052 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e27MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 145Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub047 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e29MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub047 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e29MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub066 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e21MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub122 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e7MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub078 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e18MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub078 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e18MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub080 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e17MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub070 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e21MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 145Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub093 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e15MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eY\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub043 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e34MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub108 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e10MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e2048Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 85Hz, 60\u0026micro;s. Contact 2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"9\" rowspan=\"10\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub051 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e36MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub051 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e36MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub064 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e24MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 1\u0026ndash;3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub080 M1S0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e24MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003esub080 M1S1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e24MFU\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e4000Hz\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eLSTN: 1mA, 125Hz, 60\u0026micro;s. Contact 1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003eRSTN: 0\u0026ndash;2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003eMXSX: medication and stimulation status: 1 for ON, 0 for OFF. MFU (Months Follow-Up): months since deep brain stimulation surgery. LSTN/RSTN: Left and Right Subthalamic Nucleus. Y/N: Yes/No. Recording contacts in the STN are always positioned with one above/one below the stimulation contact(s).\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eStimulation artifacts are reproducible across patients\u003c/h2\u003e\u003cp\u003eIn order to generate synchronization pulses, two abrupt changes of stimulation amplitude are performed at the beginning and at the end of each recording session by sliding the amplitude up and down quickly (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea, see methods for detailed protocol). These changes in DBS amplitude gave rise to easily identifiable artifacts in both the Percept-LFP data and in an externally recorded bipolar channel placed around the IPG and the cable (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). In intracranial recordings, the Percept\u0026trade; hardware suppresses DBS artifacts after a first visible deflection in the signal. In the external bipolar electrode, the artifact is continuously recorded when DBS is turned on and is seen as a fast, repetitive change in signal amplitude, matching the stimulation frequency (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea, \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). While the artifact shape in intracranial recordings can be slightly different, it is stable in external recordings even with various sampling frequencies (2048 Hz, 4kHz, 4096Hz). In both signals the polarity of the artifact can vary but this is accounted for automatically in DBSsync (correct detection of polarity: 100% in both signal types).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eDBSsync enables automatic and reliable synchronization of multimodal data\u003c/h3\u003e\n\u003cp\u003eOut of 34 recording sessions, automatic detection of the synchronization artifact was successful in 34 external bipolar channels (100%) and in 33 Percept-LFP channels (97%) (see Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb for an example of correct artifact detection). In the remaining Percept-LFP recording, the start of the artifact timepoint was detected two samples too early and was therefore adjusted manually. As described in the methods section, cardiac artifacts were used to determine temporal accuracy of our synchronization method (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec). Using the best sample for synchronization in 8 independent STN (i.e. not recorded during the same session), cardiac artifacts after 10s of recording were never shifted by more than 6ms between Percept-LFP and external ECG, therefore achieving a temporal resolution of 8ms (6ms corresponding to a difference of 1 to 2 samples in the Percept-LFP data sampled at 250Hz).\u003c/p\u003e\n\u003ch3\u003eDBSsync helps detect and correct timescale inaccuracies\u003c/h3\u003e\n\u003cp\u003eAs DBS pulses were repeated at the end of each recording session, the timestamps of the last artifact in the intracranial and external recordings were systematically compared after synchronization to assess data integrity and potential timeshift. As described in the methods section, timeshift may arise after longer recording duration due to small variations in the sampling frequency of the Percept\u0026trade; device around the 250Hz value indicated and, in less common cases, from the result of packet loss. DBSsync automatically detects packet loss when the .JSON file is directly imported, and adds NaN values in place of the missing data. Out of 34 recording sessions, 3 contained packet loss and were flagged by DBSsync as \u0026ldquo;corrected for packet loss\u0026rdquo;. Positive or negative timeshift values were found in all recordings (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;std: 11\u0026thinsp;\u0026plusmn;\u0026thinsp;26ms). For all recording sessions, the \u003cem\u003eeffective sampling frequency\u003c/em\u003e was computed and varied between 249.985Hz and 250.024Hz (mean\u0026thinsp;\u0026plusmn;\u0026thinsp;std: 249.997\u0026thinsp;\u0026plusmn;\u0026thinsp;0.008Hz), which is a reasonable range given the small variations expected. Therefore, this computed \u003cem\u003eeffective sampling frequency\u003c/em\u003e was used to create a more accurate timescale for Percept-LFP recordings and synchronize it again with external data, ensuring higher accuracy and consistent alignment of the signals over longer periods of time. After this adjustment of the sampling frequency, timeshift at the end of each recording was computed again and was always close to 0ms (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed).\u003c/p\u003e\n\u003ch3\u003eDBSsync enables reliable cleaning of Percept-LFP data from cardiac artifacts\u003c/h3\u003e\n\u003cp\u003e30 out of 68 STN contained visible cardiac artifacts. For each of these recordings, R-peaks were automatically detected, and results were compared with visual verification across 1min of recording (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). In 2 STN from dataset 1, R-peaks were of small amplitude compared to true brain signal and could not be detected properly because of the absence of an external ECG channel. These 2 STN recordings were therefore not included in subsequent analysis. For each of the 28 remaining STN the best possible method was used (either solely the Percept-LFP data or in combination with a synchronized external ECG channel when available). Across sessions, DBSsync detected overall 97.2% of true positive R-peaks, 0.4% of false positive and 2.4% of R-peaks were missed (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea, see black arrows for missed peaks and grey dots for correctly detected R-peaks). Each of the three cleaning methods was then applied separately and an overlap of raw and cleaned signals was plotted in the GUI (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb shows an example of the result for each method and their associated power spectrums). Power spectrums of raw and cleaned channels were also saved for subsequent analysis and evaluation of each cleaning method. A one-way ANOVA was performed to compare the effect of these three different cleaning methods on ECG suppression ratio. The test revealed that there was a statistically significant difference in ECG suppression ratio between at least two cleaning methods (F (2, 25) = [5.346], p\u0026thinsp;=\u0026thinsp;0.0067). Post hoc pairwise \u003cem\u003et\u003c/em\u003e-tests with Bonferroni correction indicated that both the template subtraction and SVD method achieved significantly greater ECG suppression compared to the interpolation method (p\u0026thinsp;=\u0026thinsp;0.0109 and p\u0026thinsp;=\u0026thinsp;0.0111, respectively). No significant difference was found between the template subtraction and SVD methods. Beta peak recovery was assessed in 7 STN presenting a prominent beta peak. There was no significant difference in beta peak recovery across methods, with all methods achieving values close to 1 indicating good preservation of beta oscillatory activity without spectral enhancement (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eIn this paper, we present a method to perform temporally precise offline synchronization of intracranial with external recordings using our toolbox DBSsync. One of the major achievements of this toolbox is multimodal data synchronization of at least 5 different signal types (EEG, kinematics, task-related events, physiological recordings (EOG, ECG) and audio) with Percept-LFP data. Following a precise protocol to deliver synchronization pulses, stimulation-induced artifacts were reproducible and reliable, resulting in an accurate automatic detection in most cases. Similar to earlier reports by Soh and colleagues (11), we could demonstrate a temporal precision of 8ms by relying on cardiac artifacts (a method that was more accessible in our setup than relying on Transcranial Magnetic Stimulation (TMS) pulses as they did). In their study, the temporal overlap of TMS pulses was only assessed 10 seconds after synchronization. While this is a good method to validate the accurate synchronization of data at the beginning of a recording, it doesn’t allow us to quantitatively assess the consistency and persistence of time-locking between external and Percept-LFP data across longer recording periods. A verification of the synchronization after 10-20 minutes of recording as proposed in our paradigm is highly valuable, as it is a more realistic duration for behavioral tasks with patients. This verification also revealed an important aspect that had not been addressed in implementations so far: Percept-LFP data can contain timeshift due to either packet loss (short interruptions of connection) or minor variations in the sampling frequency. A strategy to avoid interruptions is to always ensure proximity (\u0026lt;1 meter) and absence of obstacles (even the patient’s own body) between the clinician tablet and the IPG during the recording and avoid the automatic sleep mode by touching regularly the tablet screen. The repetition of synchronization pulses at the end of each recording session allows to easily detect and correct timeshifts by adding NaN values during missing periods to compensate for packet loss and recomputing the \u003cem\u003eeffective sampling frequency\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eAdditionally, DBSsync offers a solution to another unmet need in Percept-LFP data preprocessing which is a ready-to-use cardiac artifact removal pipeline. This pipeline demonstrated an accurate and precise R-peak detection, a clear reduction in artifactual low-frequency power achieved mostly by the template subtraction and the SVD methods and a good recovery of beta oscillatory activity. As the SVD method shows similar efficiency results as the template subtraction method but can better adapt to variations in the shape of the R-peak artifact, this method is the one we recommend using, therefore agreeing with the recommendations of Stam and colleagues (13).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWith respect to the recording possibilities of the Percept\u003csup\u003eTM\u003c/sup\u003e device, our paradigm can solely be used in the “BrainSense Streaming” mode, as it relies on the induction of stimulation artifacts. Thus, only one bipolar recording per hemisphere is possible with this synchronization technique. To perform synchronization with recordings from all contact pairs (within the so-called “Indefinite Streaming mode” of the Percept PC device), other methods need to be explored. In this regard, a recent paper proposed a different method for synchronization purposes, which relies on computer-driven artifact injection via triggered transcutaneous stimulation (14). This method uses a NeuroOmega neurophysiology system to synchronize intracranial data with task event markers but has not yet been used to synchronize with other external data such as accelerometers or EEG.\u003c/p\u003e\n\u003cp\u003eOur method for synchronization has been developed based on the Percept\u003csup\u003eTM\u003c/sup\u003e neurostimulator and a TMSi SAGA digital amplifier, but the concept can be adjusted to other neurostimulators with sensing capacity and other data recorders allowing for an external bipolar electrode. Thanks to the versatility of the data types that can be streamed to LSL, our method allows for a high number of possibilities regarding the multimodal acquisition of data due to its compatibility with .XDF data formats.\u003c/p\u003e\n\u003cp\u003eOverall, DBSsync is an open-source toolbox that doesn’t require a large setup or expensive equipment and enables intuitive use through a GUI. It is also versatile in terms of the various recording types that can be synchronized with Percept-LFP data as well as in its preprocessing capacities to improve data quality and temporal precision. Our toolbox therefore provides the opportunity to synchronize intracranial brain activity along with externally recorded signals of interest (cortical EEG, accelerometers, behavioral task events, hand-tracking devices, etc.), for example in the framework of standardized behavioral tasks. This tool will help us realize new research protocols with multimodal recordings integrated to gain further insights into biomarkers, pathophysiology, brain networks and treatment effects in PD.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cp\u003e\u003cstrong\u003ePatients\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwenty-five patients with PD and bilaterally implanted Medtronic SenSight leads in the STN connected to the Percept\u003csup\u003eTM\u003c/sup\u003e neurostimulator (PC/RC, Medtronic, Minneapolis, MN, USA) were included in the study. Three different dataset types coming from three different studies were used to develop and validate the toolbox (Table 1).\u003c/p\u003e\n\u003cp\u003eWritten informed consents were obtained from all patients, and the three studies were approved by the ethics committee at the Charit\u0026eacute; Universit\u0026auml;tsmedizin Berlin (Dataset 1: EA2_256_20, Datasets 2 and 3: EA1/164/23) and conducted following the standards set by the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDatasets\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eDataset 1\u003c/u\u003e including \u003cstrong\u003eaccelerometer data and LFP\u003c/strong\u003e: recordings were performed 3-24 months after DBS surgery as part of a long-term study investigating finger-tapping movements using accelerometers and intracranial LFP both with and without active DBS (n = 10 recording sessions).\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eDataset 2\u0026nbsp;\u003c/u\u003eincluding \u003cstrong\u003eEEG and LFP during a behavioral task\u003c/strong\u003e: recordings were performed 7-38 months after DBS surgery as part of a study recording cortical and intracranial LFP during cued button presses both with and without active DBS (n = 19 recording sessions).\u003c/p\u003e\n\u003cp\u003e\u003cu\u003eDataset 3\u0026nbsp;\u003c/u\u003eincluding \u003cstrong\u003ehand movements in 3D-space, audio and LFP during a behavioral task\u003c/strong\u003e: recordings were performed 24-36 months after DBS surgery as part of a study recording kinematics, audio and intracranial LFP during a behavioral task both with and without active DBS (n = 5 recording sessions).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRecording setup\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor all datasets, intracranial brain activity was streamed from bilaterally implanted DBS-electrodes with a sampling frequency of 250 Hz via a Bluetooth connection to a tablet used for clinical programming. As external recording source for the synchronization artifact, a bipolar electrode with one contact placed close to the IPG and the other placed close to the IPG cable was connected to a digital amplifier (TMSi International, Oldenzaal, NL \u0026ndash; TMSi SAGA) (Figure 1a).\u003c/p\u003e\n\u003cp\u003eIn dataset 1, two tri-axial accelerometers placed on both index fingers were recorded via the same digital amplifier. The output file of our digital amplifier was a .POLY5 file.\u003c/p\u003e\n\u003cp\u003eIn dataset 2, the digital amplifier was also connected to a 32-channels EEG cap (BrainWave Cap Infinity \u0026ndash; 32ch+gnd, Ag/AgCl \u0026ndash; electrodes), an accelerometer placed on the finger used to press a response button during a behavioral task and two more bipolar electrodes used as electrooculogram\u0026nbsp;(EOG) and ECG channels. Task events and signals from the digital amplifier were sent as two separate streams\u0026nbsp;to the Lab Recorder application using the Lab Streaming Layer framework\u0026nbsp;(LSL, 15) which allowed for online automatic synchronization of these data streams. Using the Lab Recorder application, each recording resulted in a single output file in the eXtensible Data Format (.XDF) containing all synchronized external streams (Figure 1b).\u003c/p\u003e\n\u003cp\u003eIn dataset 3, the digital amplifier was solely used for recording the bipolar electrode used for synchronization (around the IPG). LSL was also used during the experiment and received four different streams: one from the digital amplifier containing the signal from the bipolar electrode, one from the behavioral task containing the event markers, one from the Ultraleap camera (Leap Motion Controller 2 (Ultraleap Ltd., Bristol, UK, 2023) or Stereo IR 170 Camera (Ultraleap Ltd., Bristol, UK, 2020) recording hand movements in 3D space and one from the audio recorder. These four streams resulted in a single .XDF file as output.\u003c/p\u003e\n\u003cp\u003eThe output file of Percept-LFP recordings was a JavaScript Object Notation (.JSON) file, containing all recordings and metadata from an experimental session.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInduction of DBS synchronization artifacts\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe established a specific paradigm to induce stimulation artifacts in intracranial and external recordings for later data alignment. This paradigm consisted of two stimulation pulses at the beginning and end of each recording session (purple line in Figure 1a). Specifically, the ramp option was deactivated to allow for large amplitude pulses and baseline recordings were started with DBS turned on and bilaterally set to 0mA for 5s. Two short pulses (approximately 2s with an interval of 2s each) of unilateral, high-frequency DBS at 1mA amplitude were then delivered. DBS was then either kept at 0mA (DBS-Off sessions, n=17) or manually ramped up to the amplitude(s) tested in the study (DBS-On sessions, n=17) (Table 1). Before ending each recording, DBS was set back to 0mA bilaterally and two sequences of unilateral, high-frequency DBS with the same settings and in the same hemisphere as in the beginning were delivered.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMain interface and compatible file formats\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDBSsync is a Python-based open-source toolbox developed to align intracranial Percept-LFP recordings and external recordings (available at: https://github.com/juliettevivien/DBSsync.git). The toolbox is designed with a GUI for easier use, and a user guide is available in the supplementary data. Within DBSsync, a Percept-LFP recording from one session (.JSON, .MAT or .FIF file) and the corresponding external recording from the same session (.XDF, .POLY5 or .FIF file) are loaded. If the external file is .XDF, the user is first asked to select which LSL stream contains the bipolar electrode used for synchronization. MAT file compatibility was implemented and designed for users who use the open-source Perceive toolbox (available at https://github.com/neuromodulation/perceive/) to extract each recording session as a single .MAT file. FIF files compatibility was added for both external and intracranial files to increase compatibility for MNE-python users (16; MNE-python) who might have their own preprocessing pipeline for .JSON files and/or different output formats for the external recordings.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSynchronization of intracranial Percept-LFP and external recordings\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the time series of the external bipolar electrode recorded with the digital amplifier, the DBS artifact is detected as a steep and sustained decrease/increase (depending on electrode polarity) in signal amplitude after applying a high pass filter (1Hz) to detrend the data. The sample with the first highest amplitude change from baseline is selected as \u0026ldquo;start of the artifact\u0026rdquo; in the external channel (grey dashed line in Figure 2b upper plot). In the intracranial recording, the DBS artifact is a sharp change in signal amplitude of the signal followed by a slow recovery (Figure 2a). To identify at which point the sharp change occurs, a threshold window is computed based on the first two seconds of each intracranial recording. The last sample that lies within the value distribution of the window before crossing the threshold is detected as the start of the artifactual period (sample 0 in Figure 2c). As the artifact in the Percept-LFP data spreads over several samples and doesn\u0026rsquo;t have a consistent shape across patients, we used a numerical labeling of each sample after this sharp change. To test which sample should be chosen as actual \u0026ldquo;start of the artifact\u0026rdquo;, 8 STN from independent sessions containing clear cardiac artifacts in both the Percept-LFP data and in an external ECG channel were used. Each sample was successively used for synchronization and the alignment of endogenous cardiac artifacts R-peaks in intracranial and external recordings was assessed 10 seconds later to reproduce the analysis performed with TMS pulses by Soh and colleagues (11) for comparison. In all artifacts\u0026rsquo; shape, the sample who overall provided the best temporal alignment of endogenous cardiac artifacts was the 4\u003csup\u003eth\u003c/sup\u003e sample after the sharp change (black sample in Figure 2c) (mean time difference across 8 recording sessions \u0026plusmn; std: 0 \u0026plusmn; 4ms, min: -6ms, max +6ms). Therefore, the 4\u003csup\u003eth\u003c/sup\u003e sample is defined as \u0026ldquo;start of the artifact\u0026rdquo; and is always used for synchronization in DBSsync (grey dashed line in Figure 2b lower plot).\u003c/p\u003e\n\u003cp\u003eIn the GUI, a plot shows the sample automatically chosen as the start of the artifact in both time series, allowing the user to decide whether the artifact has been properly detected. If the automatic detection method fails due to an unusual artifact shape or a baseline contaminated by other artifacts, the user can correct the DBS artifact detection and manually select the correct sample timepoint.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eVerification of the consistency of synchronization over time\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the timeshift window, the two synchronized channels can be plotted and overlapped to check the consistency of the synchronization over using artifacts generated at the end of the recording sessions (Figure 2d). DBSsync offers the possibility to compute the timeshift (i.e. the time-delay between these end-artifacts, see bottom plots in Figure 2d). If the absolute value of the timeshift is high (\u0026gt;200ms), it suggests that data loss happened in the Percept-LFP data, most times due to communication issues between the IPG and the clinician tablet. Such recordings should not be analyzed further before properly handling this data loss. If the Percept-LFP data wasn\u0026rsquo;t loaded using the original .JSON file, consider using this method as it automatically corrects for packet loss. If the absolute value of the timeshift is below this value but still larger than 0, it might reflect a slight variation in the sampling frequency of the Percept-LFP data from 250Hz. This might not be considered relevant during short experiments but for longer recording sessions it produces inconsistent synchronization of the signals towards the end (up to 65 milliseconds of timeshift was found in our datasets for a recording of ~11minutes). Therefore, DBSsync also encompasses the possibility to compute the \u003cem\u003eeffective sampling frequency\u003c/em\u003e of the Percept-LFP data using the sampling frequency of the external digital amplifier as ground truth (usually much higher sampling frequency and data are not sent via Bluetooth, but via USB connection which allows for higher accuracy in sampling frequency information). Once this \u003cem\u003eeffective sampling frequency\u003c/em\u003e is computed, it is automatically added to the Percept-LFP metadata and the synchronization artifact must be detected again in the LFP channel using this new sampling frequency for better accuracy. When computing the timeshift again, this results in a value close to 0ms (Figure 2d, right lower plot).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCardiac artifact removal from Percept-LFP data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDBSsync offers the possibility to clean Percept-LFP data from cardiac artifacts either after synchronization with external data or independently. The first step is to detect the R-peaks in the LFP channel (Figure 3a). The contaminated channel can be used on its own or in combination with a synchronized external ECG channel recorded through the digital amplifier (recommended for better detection of R-peaks). A low-pass filter can be applied to the LFP data prior to cleaning, to remove potential stimulation artifacts and enhance signal-to-noise ratio.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIf the Percept-LFP data is used alone, the detection algorithm will segment the signal into overlapping 1-second windows and search for negative and positive peaks in each window to get a first approximation of the R-peaks localization and their polarity. 1-second epochs centered on these R-peaks are generated and averaged to create a first ECG template. The cross-correlation between this ECG template and the signal is calculated and R-peaks are identified if they exceed a 95\u003csup\u003eth\u003c/sup\u003e percentile threshold with a minimum distance of 0.5s between peaks. A refined ECG template is computed, and a second-pass detection is performed in which the threshold can be modified by the user to refine the detection.\u003c/p\u003e\n\u003cp\u003eIf an external ECG channel is synchronized and available to help detect R-peaks, this channel is band-pass filtered between 0.5Hz and 60Hz to remove slow drifts and potential DBS artifacts and the algorithm searches for the timestamps of R-peaks in the ECG using a 95th percentile threshold and a minimum distance of 0.5s between peaks. Corresponding R-peaks are searched in the LFP data by creating time windows from -80ms to +80ms around each R-peak timepoint as detected in the ECG (the algorithm searches for local minima/maxima values in each window in the LFP data). These two detection methods are reproduced from the description provided by Stam and colleagues (13). DBSsync offers the possibility to manually override some of the parameters for better control of R-peak detection (e.g. wrong polarity detected, artifactual periods to avoid, percentile threshold to use).\u003c/p\u003e\n\u003cp\u003eThe second step is to clean the signal from ECG artifacts based on the detected R-peaks. The methods implemented in DBSsync are also adapted from the paper published by Stam and colleagues (13), in which they extensively describe three different pipelines for cardiac artifact removal in Percept-LFP recordings. These three methods are available in DBSsync to dampen the cardiac artifact: 1) an interpolation method, in which the R-peaks are replaced by linearly interpolated data, 2) a template subtraction method, in which an average template of the QRS complex is generated, which is then fitted and subtracted to each cardiac artifact individually, and 3) a Singular Value Decomposition (SVD) method, in which the cardiac artifact is reconstructed at each R-peak based on the first \u0026ldquo;k\u0026rdquo; components of the SVD, fitted and subtracted separately, to account for differences in artifact shape. The number \u0026ldquo;k\u0026rdquo; of SVD components to use for the reconstruction is set by the user, based on the shape (as described in 13) and explained variance of each component.\u003c/p\u003e\n\u003cp\u003eDBSsync automatically displays four output plots to assess the quality of cardiac artifact removal: 1) detected R-peaks (Figure 3a), 2) ECG artifact shape (average template across epochs or average reconstruction using SVD components), 3) an overlap of the raw and cleaned channel (Figure 3b, left) and 4) an overlap of the power spectrum of the raw and cleaned data (Figure 3b, right). Once cardiac artifacts are properly removed, the user can choose to replace the original channel with the cleaned one. The original channels will still be saved in the output file but labeled as \u0026ldquo;RAW\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eValidation of cardiac artifact removal\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAcross 68 STN recorded, 30 contained visible cardiac artifacts (Table 1). R-peak detection was performed in these STN recordings without using any external ECG channel and the accuracy of the detection was assessed by counting true positive, false positive and missed peaks over one minute of recording (Figure 3a). The minute chosen for testing R-peak detection accuracy was always from 80 to 140s, as it was the first common minute free of synchronization pulses in all recordings. When available, this was performed again using an external ECG channel to help R-peaks detection. The best detection was kept for further cleaning of the channel. In some cases, the \u0026quot;manual override\u0026quot; option was used to correct for ECG artifact polarity (5 cases) or the start/end of detection (to avoid stimulation pulses, which can bias the ECG artifact shape if included in the template).\u003c/p\u003e\n\u003cp\u003ePower spectrums of raw and cleaned channels were computed using Welch\u0026rsquo;s method with a 1s window and 50% overlap. These power spectrums were used to determine ECG suppression ratio and beta peak recovery achieved by each method. These metrics were chosen and adapted from the thesis of Silvi L. (17). The ECG suppression ratio measures power attenuation with higher values indicating greater cardiac artifact removal. It is calculated as:\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/p\u003e\n\u003cp\u003eThe peak prominence is calculated relative to the surrounding spectral baseline within an extended frequency window (\u0026plusmn;5 Hz around peak frequency). In our datasets, 7 STN contained a prominent beta peak and were used for this analysis.\u003c/p\u003e\n\u003cp\u003eFor statistical assessment of each metric, a one-way ANOVA was performed to evaluate overall differences between the three cleaning methods, followed by Bonferroni-corrected post-hoc pairwise \u003cem\u003et\u003c/em\u003e-tests to identify specific significant differences between methods.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSaving synchronized and/or preprocessed recordings\u003cbr\u003e\u003c/strong\u003eThe saving option can be set by the user in the config file before starting DBSsync. Available saving options are .SET, .FIF, .MAT or .PKL. In the case of external .XDF files, users should be cautious of what type of stream was contained in it: if it only contained continuous streams and marker streams, the synchronized/cleaned data can be saved in .SET, .MAT or .FIF format. However, if the .XDF file contained discontinuous streams (i.e. without a defined sampling frequency but only discrete data points) the synchronized data must be saved as .PKL (as the other formats only allow for one common sampling frequency for all channels).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eDATA AVAILABILITY\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData are available conditionally through data-sharing agreements in accordance with data privacy statements signed by the patients within the legal framework of the General Data Protection Regulation of the European Union within a time frame of 6\u0026thinsp;months. Requests should be directed to A.A.K. ([email protected]) or the Open Data officer ([email protected]).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCODE AVAILABILITY\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe code of the toolbox is publicly available in a Github repository and can be accessed via this link: https://github.com/juliettevivien/DBSsync\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eACKNOWLEDGEMENTS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe would like to thank our patients for their participation in our research protocols. We would also like to thank Jennifer K. Behnke, Varvara Mathiopoulou, Johannes L. Busch and Aicha Ben Jannet who contributed in part to data collection from dataset 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFUNDING DECLARATION\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eJV has received a Doctoral Research Grant from the German Academic Exchange Service \u0026ndash; Deutscher Akademischer Austauschdienst (DAAD). RL and JGVH are participants in the BIH Charit\u0026eacute; Clinician Scientist Program funded by the Charit\u0026eacute; \u0026ndash; Universit\u0026auml;tsmedizin Berlin, and the Berlin Institute of Health at Charit\u0026eacute; (BIH). This work was supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) \u0026ndash; Project-ID 424778381 \u0026ndash; TRR 295; and by the Lundbeck Foundation as part of the collaborative project grant ADAPT-PD (Grant Nr. R336-2020-1035).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAUTHOR CONTRIBUTIONS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eJV\u0026nbsp;\u003c/strong\u003ehas designed the methodology, performed experiments, collected and analyzed data, developed the software and graphical user interface, wrote original drafts and made revisions, created scientific figures and acquired funding.\u0026nbsp;\u003cstrong\u003eCES, LH, MH, AM and LKF\u0026nbsp;\u003c/strong\u003ehave performed experiments and collected data.\u0026nbsp;\u003cstrong\u003eCES\u003c/strong\u003e also participated in data analysis, tested code components and reviewed and edited the original draft of the manuscript.\u0026nbsp;\u003cstrong\u003eLKF\u003c/strong\u003e and\u0026nbsp;\u003cstrong\u003eRL\u003c/strong\u003e were involved in conceptualization and provided critical review and editing to the manuscript.\u0026nbsp;\u003cstrong\u003eJGVH\u0026nbsp;\u003c/strong\u003eand \u003cstrong\u003eAC\u0026nbsp;\u003c/strong\u003ehavecontributed to the programming and testing of existing code components. \u003cstrong\u003eAAK\u0026nbsp;\u003c/strong\u003ewas involved in conceptualization, critical review and editing of the manuscript, supervision and provided resources and funding. All authors read and approved of the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCOMPETING INTERESTS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.A.K. has served on advisory boards of Medtronic and has received honoraria and travel support from Medtronic, Boston Scientific, Ipsen Pharma and Teva.\u003c/p\u003e\n\u003cp\u003eL.K.F. has received speaker honoraria from Medtronic.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eDeuschl, G., et al. A randomized trial of deep-brain stimulation for Parkinson\u0026apos;s disease. The New England journal of medicine, 355(9), 896\u0026ndash;908 (2006).\u003c/li\u003e\n\u003cli\u003eVolkmann, J. et al. Pallidal neurostimulation in patients with medication-refractory cervical dystonia: a randomised, sham-controlled trial. The Lancet Neurology 13, 875\u0026ndash;884 (2014).\u003c/li\u003e\n\u003cli\u003eVan Rheede, J. J. et al. Diurnal modulation of subthalamic beta oscillatory power in Parkinson\u0026rsquo;s disease patients during deep brain stimulation. npj Parkinsons Dis. 8, 88 (2022).\u003c/li\u003e\n\u003cli\u003eMathiopoulou, V. et al. Gamma entrainment induced by deep brain stimulation as a biomarker for motor improvement with neuromodulation. Nat Commun 16, 2956 (2025).\u003c/li\u003e\n\u003cli\u003eOlaru, M. et al. Deep brain stimulation-entrained gamma oscillations in chronic home recordings in Parkinson\u0026rsquo;s disease. Brain Stimulation 18, 132\u0026ndash;141 (2025).\u003c/li\u003e\n\u003cli\u003eFarokhniaee, A. et al. Gait-related beta-gamma phase amplitude coupling in the subthalamic nucleus of parkinsonian patients. Sci Rep 14, 6674 (2024).\u003c/li\u003e\n\u003cli\u003eLouie, K. H., Balakid, J. P., Azgomi, H. F., Marks, J. H. \u0026amp; Wang, D. D. Adaptive deep brain stimulation timed to gait phase improves walking in Parkinson\u0026rsquo;s disease. medRxiv : the preprint server for health sciences (2025)\u003c/li\u003e\n\u003cli\u003eBusch, J. L. \u0026amp; Kaplan, J. et al. Chronic adaptive deep brain stimulation for Parkinson\u0026rsquo;s disease: clinical outcomes and programming strategies. npj Parkinsons Dis. 11, 264 (2025).\u003c/li\u003e\n\u003cli\u003eSwinnen, B. E. K. S., et al. Basal ganglia theta power indexes trait anxiety in people with Parkinson\u0026apos;s disease. Brain: a journal of neurology, 148(4), 1228\u0026ndash;1241 (2025).\u003c/li\u003e\n\u003cli\u003eThenaisie, Y. et al. Towards adaptive deep brain stimulation: clinical and technical notes on a novel commercial device for chronic brain sensing. J. Neural Eng. 18, 042002 (2021).\u003c/li\u003e\n\u003cli\u003eSoh, C., Hervault, M., Rohl, A. H., Greenlee, J. D. W. \u0026amp; Wessel, J. R. Precisely-timed outpatient recordings of subcortical local field potentials from wireless streaming-capable deep-brain stimulators: a method and toolbox. Journal of Neuroscience Methods 418, 110448 (2025).\u003c/li\u003e\n\u003cli\u003eHammer, L. H., Kochanski, R. B., Starr, P. A. \u0026amp; Little, S. Artifact Characterization and a Multipurpose Template-Based Offline Removal Solution for a Sensing-Enabled Deep Brain Stimulation Device. Stereotact Funct Neurosurg 100, 168\u0026ndash;183 (2022).\u003c/li\u003e\n\u003cli\u003eStam, M. J. et al. A comparison of methods to suppress electrocardiographic artifacts in local field potential recordings. Clinical Neurophysiology 146, 147\u0026ndash;161 (2023).\u003c/li\u003e\n\u003cli\u003eAlarie, M. E., Provenza, N. R., Herron, J. A. \u0026amp; Asaad, W. F. Automated artifact injection into sensing-capable brain modulation devices for neural-behavioral synchronization and the influence of device state. Brain Stimulation 16, 1358\u0026ndash;1360 (2023).\u003c/li\u003e\n\u003cli\u003eKothe, C. et al. The lab streaming layer for synchronized multimodal recording. Imaging Neuroscience 3, IMAG.a.136 (2025).\u003c/li\u003e\n\u003cli\u003eGramfort, A. MEG and EEG data analysis with MNE-Python. Front. Neurosci. 7, (2013).\u003c/li\u003e\n\u003cli\u003eSilvi, L. Electroencephalography and Local Field Potential fusion to characterize Deep Brain Stimulation in Parkinson\u0026rsquo;s disease. [Master\u0026rsquo;s thesis] (2025)\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"npj-parkinsons-disease","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjparkd","sideBox":"Learn more about [npj Parkinson's Disease](http://www.nature.com/npjparkd/)","snPcode":"41531","submissionUrl":"https://submission.springernature.com/new-submission/41531/3","title":"npj Parkinson's Disease","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8228751/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8228751/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eImplanted deep brain stimulation devices are now capable of chronically recording activity from intracranial brain areas during stimulation. This new type of data has the potential to increase our understanding of disease-related brain activity and its modulation in response to therapy or other types of stimuli. With the innovative approach of adaptive deep brain stimulation now clinically available, multimodal characterization of neural biomarkers becomes of utmost importance to define optimal feedback signals for adaptive brain stimulation and allow for better fine-tuning of stimulation parameters. To investigate these biomarkers, we developed DBSsync, a paradigm and an open-source Python toolbox with its graphical user interface for temporally precise synchronization of intracranial recordings with external data, allowing for multimodal research protocols. DBSsync achieves a temporal precision of 8 milliseconds and incorporates cardiac artifact removal methods to facilitate intracranial data preprocessing, thus enabling the integration and precise synchronization of multiple brain signals with external sensors and various behavioral timeline data.\u003c/p\u003e","manuscriptTitle":"DBSsync: combining intracranial and multimodal data to investigate new biomarkers in Parkinson’s disease","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-08 12:03:17","doi":"10.21203/rs.3.rs-8228751/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-01T21:47:28+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-27T01:05:44+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-16T08:48:43+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"39187278890043739581455098165962621782","date":"2026-01-05T13:11:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"214997714181207706050582696100083683959","date":"2026-01-05T12:14:57+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"146777706985086942979956426147759993003","date":"2026-01-05T09:28:29+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-12-04T08:48:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-12-02T20:24:14+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-12-02T14:48:17+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Parkinson's Disease","date":"2025-11-28T09:05:26+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-parkinsons-disease","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjparkd","sideBox":"Learn more about [npj Parkinson's Disease](http://www.nature.com/npjparkd/)","snPcode":"41531","submissionUrl":"https://submission.springernature.com/new-submission/41531/3","title":"npj Parkinson's Disease","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"914f4dbc-cd47-4a7b-a14f-cebaf2c2ac2c","owner":[],"postedDate":"December 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":59274233,"name":"Biological sciences/Biological techniques"},{"id":59274234,"name":"Health sciences/Biomarkers"},{"id":59274235,"name":"Physical sciences/Engineering"},{"id":59274236,"name":"Health sciences/Neurology"},{"id":59274237,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2026-05-19T09:38:21+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-08 12:03:17","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8228751","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8228751","identity":"rs-8228751","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}