Background: Electrode application for EEG prognostication in postanoxic coma patients is labor-intensive. We studied the usability of a self-adhesive forehead EEG electrode (Bittium BrainStatus™) and compared the accuracy of background classification with conventional EEG. Methods: In 51 postanoxic coma patients, simultaneous monitoring using conventional EEG and a forehead electrode was performed. 5-min EEG fragments were classified according to the ACNS criteria by four experts, based on the conventional montage, the forehead electrode, and a montage based on the standard EEG recording but visualized as the forehead electrode. Results: Forehead electrode recordings were of sufficient quality in 74.1% of fragments. Agreement was moderate for standard EEG versus the forehead electrode (κ: 0.56), and near perfect for conventional EEG versus conventional EEG visualized as the forehead electrode (κ: 0.85). Due to higher noise levels, detection rates of discontinuous and suppressed patterns were reduced. Conclusion: EEG forehead electrode recordings were interpretable in a large proportion of cases. If signal quality is insufficient for assessment, additional standard EEG should be performed to allow definitive assessment. Significance: The use of a self-adhesive forehead electrode set has the potential to substantially reduce EEG technician workload in the ICU and suggests considerable potential benefit as a screening-oriented EEG monitoring strategy in postanoxic coma.

Robotic systems assess joint dynamics objectively by perturbing the limb and estimating properties such as impedance. Position perturbations constrain the limb to a target trajectory, reducing variability in task execution but obstructing voluntary motion. Force perturbations allow voluntary movement but elicit orientation-dependent responses, increasing the number of trials needed for accurate estimates. To overcome these limitations, we combined the flexibility of admittance control with the repeatability of position perturbations. A minimum-jerk trajectory ensures smooth transitions. The experiment with six healthy participants was performed to demonstrate the reliability, accuracy and smoothness of applying such perturbations during voluntary movement. Reliability was the proportion of perturbations that reached the target velocity within one millisecond of the acceleration time window. Accuracy was measured as the RMSE between the target and measured velocity during the constant velocity. Smoothness was assessed as perceivability: the fraction of trials in which participants correctly detected a perturbation. The controller allows continuous voluntary movement, switching only during perturbations to impose a precise, specified perturbation. All perturbations reached the target velocity within one millisecond of the acceleration time window; thus, the method is reliable. Under the most demanding condition— an increase to 200 deg/s in 0.01 s—the RMSE between target and measured velocity was 1.1 deg/s (0.55%), indicating a high accuracy. Specially designed perturbations had a perceivability accuracy of 22.1%, indicating smooth transitions between control modes. Together, these results indicate a promising approach for assessing joint dynamics during voluntary elbow movement, enabling assessment during activities of daily living.

To increase the quality of life of stroke patients, better diagnostics with the ability to identify the cause of motor impairment are needed. Robotic diagnostics increases the resolution of measurements, allows for tracking progress over a longer period, and can be used to evaluate new treatments. The Shoulder Elbow Perturbator (SEP) was developed to improve the diagnostics of post-stroke motor impairment. The SEP has already been tested on patients, showing promising results in identifying the cause of motor impairment, but no SEP system performance analysis has been published. To identify the joint properties of the elbow accurately, the SEP should have a bandwidth of at least 12 Hz. Furthermore, admittance and velocity control are required for various possible experimental tasks. This paper shows that the SEP performs adequately for the desired perturbations and experimental conditions for system identification of the human elbow. The SEP's performance is analysed with multisine signals to determine the bandwidth and endpoint dynamics. The velocity controller bandwidth is 50 Hz, and the admittance controller bandwidth is 65 Hz. Furthermore, the controller is stable. Thus, the SEP meets all the requirements and should be able to provide the desired perturbations and experimental conditions needed for system identification of the human elbow.

Background: New acute and preventive treatments have expanded migraine care options, highlighting the need for integrated, personalized management strategies. Telemedicine and telemonitoring support multidisciplinary approaches and are essential tools in optimizing patient outcomes. Objectives: This study aims to demonstrate the usability and user-friendliness of a validated E-diary and a self-administered electroencephalogram (EEG) telemonitoring setup for migraine care and research. Methods: E-diary data were collected from adult migraine patients at the Leiden Headache Center to assess compliance, with patient satisfaction evaluated through questionnaires. In a separate component of the study, the user-friendliness of a home-based EEG setup for migraine research was examined. Participants completed two measurement sessions on different days, with varying intervals between sessions. Evaluation measures included the System Usability Scale (SUS), task completion time, electrode connection success, and overall user experience. Results: Migraine patients (n = 753) were followed for a median of 353 [IQR 128–697] days. Compliance was 96.7 % [IQR 88.1–99.6]. The E-diary received a median score of 7/10, 66.0 % of patients reported being (very) satisfied with the E-diary app. The EEG setup was tested by 20 participants and awarded a high SUS-score of 91.2 [IQR 86.2, 95.0]. Conclusion: Telemedicine and telemonitoring offer scalable, effective solutions for advancing both migraine care and research. Telemedicine with the E-diary may enhance personalized, integrated migraine care. Compliance and satisfaction with the E-diary are high. Self-administered telemonitoring using remote EEG setups demonstrates the feasibility of conducting complex studies in home-based settings.

Robotic rehabilitation systems may benefit from haptic rendering to provide sensorimotor training to patients with acquired brain injuries. Haptic rendering usually involves modulating stiffness and viscosity to simulate real-world hand-object interactions. Yet, the effect of rendering different viscosities on brain activity remains mainly unexplored. To fill this gap, we ran an experiment with twelve unimpaired participants who were asked to grasp and release virtual liquid dispensers whose stiffness and viscosity were rendered using a haptic hand rehabilitation robot. All liquid dispensers had identical wall stiffness but contained liquids of three different viscosities. We also incorporated control conditions without viscosity and stiffness rendering, involving both passive and active grasping movements. Electroencephalography data were recorded during the experiment. We found stronger ipsilateral somatosensory mu and beta event-related desynchronization during movements with viscosity and stiffness rendering compared to the control conditions, while different viscosity levels did not result in significant variations. Furthermore, no significant electroencephalography activity differences were found between control conditions. These findings indicate that while viscosity and stiffness rendering strengthens brain activity, modulating viscosity levels does not significantly affect this response. This insight may contribute to the design of rehabilitation games by informing the choice of viscosity rendering parameters.

Prognostication after moderate-to-severe traumatic brain injury (TBI) remains challenging in the intensive care unit (ICU) despite the existence of well-validated online prognostication tools. Changes in brain activity related to TBI can be measured using electroencephalography (EEG), making it a potentially interesting diagnostic tool to refine prognostication. The primary objective of this systematic review was to evaluate the literature concerning the prognostic value of EEG among patients with TBI in the ICU. Five databases were searched from inception until August 13, 2024. The search identified 1492 unique records. Eventually, 27 manuscripts met the inclusion criteria (>18 years old, Glasgow Coma Scale ≤12, EEG performed in the ICU). The QUIPS (QUality In Prognostic Studies) and PROBAST (Prediction model Risk Of Bias ASsessment Tool) tools were used to assess the study quality and bias. Due to high heterogeneity in EEG feature and outcome definitions and a lack of correction for confounding factors, all studies had a moderate-to-high risk of bias. Nonetheless, specific EEG features (identified through visual and quantitative EEG, EEG reactivity, and machine learning techniques) were found to be predictive of neurological outcomes up to 1.5 years after TBI. While epileptiform discharges and seizures were not consistently associated with outcomes, a higher alpha variability, a more continuous EEG, present EEG reactivity, and present EEG sleep features were predictive of better outcomes. The combination of EEG features with clinical parameters demonstrated improved predictive performance compared with models using standard clinical parameters alone. Still, the EEG features described and their potential additional value in outcome prediction after TBI merit further investigation.

Objective: Quantitative markers of cortical excitability may help identify responders to anti-seizure medications (ASMs). We studied the relationship between ASM load and two electroencephalography (EEG) markers of cortical excitability in people with refractory epilepsy. Methods: We included individuals with refractory focal epilepsy undergoing presurgical evaluation, involving ASM tapering and sleep deprivation. We obtained daily resting state EEG and EEG responses to visual stimulation at linearly increasing flash frequency (10–40 Hz chirp). We extracted the aperiodic exponent from resting state EEG power spectra and analysed chirp response at driving and second-order harmonic frequencies. We modelled ASM load, which we related to the EEG markers using linear mixed-effects regression. Results: Forty-eight subjects (median age 34 years, age range 16–62 years, 19 females) participated. The spectral exponent became less negative with ASM load reduction (p = 0.02), mainly attributable to reduced low-frequency power. Lowering ASM load increased the harmonic response to chirp stimulation (p = 0.004), also after accounting for sleep deprivation (p = 0.02), but did not affect the driving response. ASM tapering specifically increased harmonic responses to high stimulation frequencies (27–40 Hz, p = 0.01). Interpretation: Resting state EEG spectral exponents and visual chirp responses reflect ASM load in refractory epilepsy. Low-frequency spectral changes in resting state EEG may only mirror ASM-induced spectral slowing. Visual chirp stimulation reveals enhanced harmonic EEG responses during low ASM loads, likely due to both increased high gamma activity and increased response to visual perturbations. Implementation of the markers would need normative values to reduce the delay to individually optimised treatment regimens.

Humans vary the stiffness in their joints depending on tasks and circumstances. For posture control a high joint stiffness is required to withstand perturbations, whereas for force control a low joint stiffness is required. To investigate how humans vary their joint stiffness precisely for moving an arm, a wearable device is needed that can generate small force perturbations at the wrist while measuring the resulting muscular reactions. The majority of the state-of-the-art devices either offer too little versatility or impede the free movement of the arm. Based on a 3-DoF spatial redundant 4-RUU parallel manipulator applied in an inverted way where the original base with actuators has become the moving platform and the original moving platform is attached to the wrist as a bracelet, a versatile, 0.175 kg lightweight, low impedance, and compact wearable device was developed that can generate perturbation forces in X-, Y-, and Z-direction. The design and a prototype of the device are presented with experimental tests showing controlled perturbations in the order of 4 N with frequencies up to 12 Hz.

Background: Upper limb impairments in a hemiparetic arm are clinically quantified by well-established clinical scales, known to suffer poor validity, reliability, and sensitivity. Alternatively, robotics can assess motor impairments by characterizing joint dynamics through system identification. In this study, we establish the merits of quantifying abnormal synergy, spasticity, and changes in joint viscoelasticity using system identification, evaluating (1) feasibility and quality of parametric estimates, (2) test–retest reliability, (3) differences between healthy controls and patients with upper limb impairments, and (4) construct validity. Methods: Forty-five healthy controls, twenty-nine stroke patients, and twenty cerebral palsy patients participated. Participants were seated with the affected arm immobilized in the Shoulder-Elbow-Perturbator (SEP). The SEP is a one-degree-of-freedom perturbator that enables applying torque perturbations to the elbow while providing varying amounts of weight support to the human arm. Participants performed either a ‘do not intervene’ or a resist task. Elbow joint admittance was quantified and used to extract elbow viscosity and stiffness. Fifty-four of the participants performed two sessions to establish the test–retest reliability of the parameters. Construct validity was assessed by correlating system identification parameters to parameters extracted using a SEP protocol that objectifies current clinical scales (Re-Arm protocol). Results: Feasibility was confirmed by all participants successfully completing the study protocol within ~ 25 min without reporting pain or burden. The parametric estimates were good with a variance-accounted-for of ~ 80%. A fair to excellent test–retest reliability was found (ICC= 0.46 - 0.98) for patients, except for elbow stiffness with full weight support (ICC= 0.35). Compared to healthy controls, patients had a higher elbow viscosity and stiffness during the ‘do not intervene’ task and lower viscosity and stiffness during the resist task. Construct validity was confirmed by a significant (all p< 0.03) but weak to moderate (r= 0.36 - 0.50) correlation with parameters from the Re-Arm protocol. Conclusions: This work demonstrates that system identification is feasible and reliable for quantifying upper limb motor impairments. Validity was confirmed by differences between patients and controls and correlations with other measurements, but further work is required to optimize the experimental protocol and establish clinical value.

Careful control of joint impedance, or dynamic joint stiffness, is crucial for successful performance of movement. Time-varying system identification (TV-SysID) enables quantification of joint impedance during movement. Several TV-SysID methods exist, but have never been systematically compared. Here, we simulate time-varying joint behavior and propose three performance metrics that enable to quantify and compare TV-SysID methods. Time-varying joint stiffness is simulated using a square wave and subsequently estimated with three TV-SysID methods: the ensemble, short data segment, and basis impulse response function method. These methods were compared based on (1) bias with respect to the simulated joint stiffness, (2) random error across 100 simulation trials, and (3) maximum adaptation speed in joint stiffness that can be captured. This approach revealed that each TV-SysID method has its own unique properties. The simulation method and performance metrics pave the way for developing a framework to quantify the strengths and weaknesses of TV-SysID algorithms for estimating joint impedance.

Accurate and swift tuning of joint impedance is crucial to perform movement and interaction with our environment. Time-varying system identification enables quantification of joint impedance during movement. Many methods have been developed over the years, each with their own mathematical approach and underlying assumptions. Yet, for the identification of joint impedance, a systematic comparison revealing each method's unique strengths and weaknesses, is lacking. Here, we propose a quantitative framework to compare these methods. The framework is used to review five time-varying system identification methods using both simulated data and experimental data. These methods included three time-domain methods: ensemble, short data segment, and basis impulse response function; and two frequency-domain methods: ensemble spectral, and kernel-based regression. In the simulation study, joint stiffness – the static component of impedance – was simulated as a square wave to mimic the most extreme case of time-varying behavior. The identification results were compared based on the (1) variance accounted for (VAF), (2) bias, (3) random, and (4) total estimation error with respect to the simulated joint stiffness; and (5) rise time between two stiffness levels. In the experimental study, human ankle joint impedance was identified. Identification performance was compared using the variability in estimating joint stiffness – representative of the random error – and VAF. The performance metrics revealed distinct identification properties for each method. Therefore, researchers must make a well-justified decision which method is most appropriate for their application. The combination of simulation and experimental work with extensive performance quantification creates a framework for quantitative assessment of newly developed time-varying system identification methods.

Background: Proprioception is important for regaining motor function in the paretic upper extremity after stroke. However, clinical assessments of proprioception are subjective and require verbal responses from the patient to applied proprioceptive stimuli. Cortical responses evoked by robotic wrist perturbations and measured by electroencephalography (EEG) may be an objective method to support current clinical assessments of proprioception. Objective: To establish whether evoked cortical responses reflect proprioceptive deficits as assessed by clinical scales and whether they predict upper extremity motor function at 26 weeks after stroke. Methods: Thirty-one patients with stroke were included. In week 1, 3, 5, 12, and 26 after stroke, the upper extremity sections of the Erasmus modified Nottingham Sensory Assessment (EmNSA-UE) and the Fugl-Meyer Motor Assessment (FM-UE) and the EEG responses (64 channels) to robotic wrist perturbations were measured. The extent to which proprioceptive input was conveyed to the affected hemisphere was estimated by the signal-to-noise ratio (SNR) of the evoked response. The relationships between SNR and EmNSA-UE as well as SNR and time after stroke were investigated using linear regression. Receiver-operating-characteristic curves were used to compare the predictive values of SNR and EmNSA-UE for predicting whether patients regained some selective motor control (FM-UE > 22) or whether they could only move their paretic upper extremity within basic limb synergies (FM-UE ≤ 22) at 26 weeks after stroke. Results: Patients (N = 7) with impaired proprioception (EmNSA-UE proprioception score < 8) had significantly smaller SNR than patients with unimpaired proprioception (N = 24) [EmNSA-UE proprioception score = 8, t(29) = 2.36, p = 0.03]. No significant effect of time after stroke on SNR was observed. Furthermore, there was no significant difference in the predictive value between EmNSA-UE and SNR for predicting motor function at 26 weeks after stroke. Conclusion: The SNR of the evoked cortical response does not significantly change as a function of time after stroke and differs between patients with clinically assessed impaired and unimpaired proprioception, suggesting that SNR reflects persistent damage to proprioceptive pathways. A similar predictive value with respect to EmNSA-UE suggests that SNR may be used as an objective predictor next to clinical sensory assessments for predicting motor function at 26 weeks after stroke.

Motor learning has been linked with increases in corticospinal excitability (CSE). However, the robustness of this link is unclear. In this study, changes in CSE associated with learning a visuomotor tracking task were mapped using transcranial magnetic stimulation (TMS). TMS maps were obtained before and after training with the first dorsal interosseous (FDI) of the dominant and nondominant hand, and for a distal (FDI) and proximal (biceps brachii) muscle. Tracking performance improved following 20 min of visuomotor training, while map area was unaffected. Large individual differences were observed with 18%–36% of the participants revealing an increase in TMS map area. This result highlights the complex relationship between motor learning and use-dependent plasticity of the motor cortex.

Background: A trend in the non-invasive brain stimulation literature is to assess the outcome of an intervention using a responder analysis whereby participants are di- or trichotomised in order that they may be classified as either responders or non-responders. Objective: Examine the extent of the Type I error in motor evoked potential (MEP) data subjected to responder analyses. Methods: Seven sets of 30 MEPs were recorded from the first dorsal interosseous muscle in 52 healthy volunteers. Four classification techniques were used to classify the participants as responders or non-responders: (1) the two-step cluster analysis, (2) dichotomised thresholding, (3) relative method and (4) baseline variance method. Results: Despite the lack of any intervention, a significant number of participants were classified as responders (21–71%). Conclusion: This study highlights the very large Type I error associated with dichotomising continuous variables such as the TMS MEP.

Background During Transcranial Magnetic Stimulation (TMS) experiments researchers often use a neuronavigation system to precisely and accurately maintain coil position and orientation. New method This study aimed to develop and validate an open-source software for TMS coil navigation. StimTrack uses an optical tracker and an intuitive user interface to facilitate the maintenance of position and orientation of any type of coil within and between sessions. Additionally, online access to navigation data is provided, hereby adding e.g. the ability to start or stop the magnetic stimulator depending on the distance to target or the variation of the orientation angles. Results StimTrack allows repeatable repositioning of the coil within 0.7 mm for translation and <1° for rotation. Stimulus-response (SR) curves obtained from 19 healthy volunteers were used to demonstrate that StimTrack can be effectively used in a typical experiment. An excellent intra and inter-session reliability (ICC > 0.9) was obtained on all parameters computed on SR curves acquired using StimTrack. Comparison with existing method StimTrack showed a target accuracy similar to that of a commercial neuronavigation system (BrainSight, Rogue Research Inc.). Indeed, small differences both in position (∼0.2 mm) and orientation (<1°) were found between the systems. These differences are negligible given the human error involved in landmarks registration. Conclusions StimTrack, available as supplementary material, is found to be a good alternative for commercial neuronavigation systems facilitating assessment changes in corticospinal excitability using TMS. StimTrack allows researchers to tailor its functionality to their specific needs, providing added value that benefits experimental procedures and improves data quality.

During movements, humans continuously regulate their joint impedance to minimize control effort and optimize performance. Joint impedance describes the relationship between a joint's position and torque acting around the joint. Joint impedance varies with joint angle and muscle activation and differs from trial-to-trial due to inherent variability in the human control system. In this paper, a dedicated time-varying system identification (SI) framework is developed involving a parametric, kernel-based regression, and nonparametric, “skirt decomposition,” SI method to monitor the time-varying joint impedance during a force task. Identification was performed on single trials and the estimators included little a priori assumptions regarding the underlying time-varying joint mechanics. During the experiments, six (human) participants used flexion of the wrist to apply a slow sinusoidal torque to the handle of a robotic manipulator, while receiving small position perturbations. Both methods revealed that the sinusoidal change in joint torque by activation of the wrist flexor muscles resulted in a sinusoidal time-varying joint stiffness and resonance frequency. A third-order differential equation allowed the parametric kernel-based estimator to explain on average 76% of the variance (range 52%-90%). The nonparametric skirt decomposition method could explain on average 84% of the variance (range 66%-91%). This paper presents a novel framework for identification of time-varying joint impedance by making use of linear time-varying models based on a single trial of data.

In this study, a nonparametric method, developed in Lataire et al. (2012), is applied to the identification of linear time-varying human joint admittance. The aim of the method, denoted Skirt Decomposition method, is to reconstruct the time-varying system function. The main contribution of the paper is to evaluate the possibilities and limitations of the method for the identification of linear time-varying human joint admittance in simulation. The proposed method delivers an estimate of linear time-varying joint admittance from a single experimental trial, provided that a multisine is used as excitation signal. The trade-off between i) the frequency resolution of the dynamics, and ii) the allowable complexity of the time variation is explored.